The DOI matches its Crossref record; 78 of the 78 references it cites were found in a registry or an index; 25 of the 33 sources its claims cite were checked for support.
Object detection using YOLO: challenges, architectural successors, datasets and applications
Open accessAbstract
Object detection is one of the predominant and challenging problems in computer vision. Over the decade, with the expeditious evolution of deep learning, researchers have extensively experimented and contributed in the performance enhancement of object detection and related tasks such as object classification, localization, and segmentation using underlying deep models. Broadly, object detectors are classified into two categories viz. two stage and single stage object detectors. Two stage detectors mainly focus on selective region proposals strategy via complex architecture; however, single stage detectors focus on all the spatial region proposals for the possible detection of objects via relatively simpler architecture in one shot. Performance of any object detector is evaluated through detection accuracy and inference time. Generally, the detection accuracy of two stage detectors outperforms single stage object detectors. However, the inference time of single stage detectors is better compared to its counterparts. Moreover, with the advent of YOLO (You Only Look Once) and its architectural successors, the detection accuracy is improving significantly and sometime it is better than two stage detectors. YOLOs are adopted in various applications majorly due to their faster inferences rather than considering detection accuracy. As an example, detection accuracies are 63.4 and 70 for YOLO and Fast-RCNN respectively, however, inference time is around 300 times faster in case of YOLO. In this paper, we present a comprehensive review of single stage object detectors specially YOLOs, regression formulation, their architecture advancements, and performance statistics. Moreover, we summarize the comparative illustration between two stage and single stage object detectors, among different versions of YOLOs, applications based on two stage detectors, and different versions of YOLOs along with the future research directions.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
46 No basis given · 13 Supported · 10 Partly supported · 6 No full text of the sources · 6 Own interpretation · 2 Sources not found · 1 Not checked yet · accuracy not yet measured
Every source the paper cites for this claim was checked and supports it.
In the paper
“Authors demonstrate the capability of shallower networks in the object detection without any explicit need of the accelerators.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v1)
What the sources say
- Supportedcites [27] Huang R, Pedoeem J, Chen C (2018) YOLO-LITE: a real-time object detection algorithm optimized for non-GPU computers. In: 2018 IEEE international conference on big data (big data), pp 2503–2510. https://doi.org/10.1109/BigData.2018.8621865 source
The YOLO-LITE Architecture section (p. 3) explicitly identifies fast object detection on non-GPU computers as a capability of shallow networks. The Conclusion (p. 6) reports that YOLO-LITE ran at 21 FPS on a non-GPU computer, supporting the claim that a GPU accelerator was not required. (LLM, conf=0.94)
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“CSPNets help in rich gradient combination with minimal computation cost [73]”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v4)
What the sources say
- Supportedcites [73] Wang CY, Mark Liao HY, Wu YH, Chen PY, Hsieh JW, Yeh IH (2020) CSPNet: a new backbone that can enhance learning capability of CNN. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, pp 390–391 source
The Introduction (p. 2) states that CSPNet was designed to achieve a richer gradient combination while reducing computation, and reports a 10–20% reduction in computation effort when applied to several architectures. This directly supports the paraphrase. (LLM, conf=0.94)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“YOLO (v2) is known as an immediate successor of YOLO (v1), utilized for detecting objects for around over 9000 categories [54].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v2)
What the sources say
- Partly supportedcites [54] Redmon J, Farhadi A (2017) YOLO9000: better, faster, stronger. In: Proceedings of the IEEE conference on computer vision and pattern recognition 2017, pp 7263–7271. https://doi.org/10.48550/arXiv.1612.08242 sourceExperimental — accuracy not yet measured
The Introduction identifies YOLOv2 as an improvement on the original YOLO detector, but attributes detection of over 9000 categories to YOLO9000. The Hierarchical classification section says YOLO9000 uses the base YOLOv2 architecture with additional joint training; it does not assign that category count to YOLOv2 itself. (LLM, conf=0.88)
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“A convolutional neural network predicts the bounding boxes as well as class probabilities for all the objects depicted in an image.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs
What the sources say
- Supportedcites [56] Redmon J, Divvala S, Girshick R, Farhadi A (2016) You only look once: unified, real-time object detection. In proceedings of the IEEE conference on computer vision and pattern recognition, pp 779-788 source
The paper’s introduction states that a single convolutional network simultaneously predicts multiple bounding boxes and class probabilities. Section 2, “Unified Detection,” describes making predictions across the image and reasoning about its objects. The claim describes YOLO’s prediction task, not a guarantee that it detects every object correctly. (LLM, conf=0.94)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“The performance of a structured model after deployment is mostly dominated by the model complexity [10].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models
What the sources say
- Partly supportedcites [10] Chen K, Franko K, Sang R (2021) Structured model pruning of convolutional networks on tensor processing units. ArXiv preprint arXiv:210704191. https://doi.org/10.48550/arXiv.2107.04191 sourceExperimental — accuracy not yet measured
The Introduction (p. 1) says deployment is constrained by model size, memory usage, and computational time, and the Results (p. 2) link pruning to improved memory usage and step time. But the paper measures performance during training, not after deployment, and does not establish that model complexity mostly dominates deployed performance. (LLM, conf=0.82)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Another variant of YOLO, known as Yolo-lite is proposed, specifically designed for real time object detection on a standard non-GPU system [27].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v1)
What the sources say
- Supportedcites [27] Huang R, Pedoeem J, Chen C (2018) YOLO-LITE: a real-time object detection algorithm optimized for non-GPU computers. In: 2018 IEEE international conference on big data (big data), pp 2503–2510. https://doi.org/10.1109/BigData.2018.8621865 source
The Introduction (p. 1) says YOLO-LITE uses YOLO as a starting point and aims for real-time object detection on a standard non-GPU computer. The Conclusion (p. 6) reports that it ran at 21 FPS on a non-GPU computer. (LLM, conf=0.95)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Mostly, localization errors are either due to occupancy of background in the predictions and detecting similar objects [45].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Introduction > Challenges in object detection
What the sources say
- Partly supportedcites [45] Liu L, Ouyang W, Wang X, Fieguth P, Chen J, Liu X, Pietikäinen M (2020) Deep learning for generic object detection: a survey. Int J Comput Vis 128(2):261–318. https://doi.org/10.48550/arXiv.1809.02165 sourceExperimental — accuracy not yet measured
Section 9, p. 24 says background pixels included in bounding boxes can affect localization accuracy. But Section 9, p. 25 treats confusion between similar objects as a separate form of error, not a cause of localization error; it instead identifies insufficient overlap and duplicate detections as possible causes. The survey does not say the claimed causes account for most localization errors. (LLM, conf=0.88)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Moreover, they also demonstrate that the presence of batch normalization limits the performance of shallow neural networks in the course of object detection.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v1)
What the sources say
- Partly supportedcites [27] Huang R, Pedoeem J, Chen C (2018) YOLO-LITE: a real-time object detection algorithm optimized for non-GPU computers. In: 2018 IEEE international conference on big data (big data), pp 2503–2510. https://doi.org/10.1109/BigData.2018.8621865 sourceExperimental — accuracy not yet measured
The YOLO-LITE Architecture section (p. 3) says batch normalization slows the overall speed of a shallow object-detection network, and the Batch Normalization section (p. 5) explains the added feedforward calculations. The claim broadens this speed-specific finding to overall performance; the paper also notes that batch normalization can improve accuracy (p. 4). (LLM, conf=0.84)
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Authors of YOLO [56] have reframed the problem of object detection as a regression problem instead of classification problem.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs
What the sources say
- Supportedcites [56] Redmon J, Divvala S, Girshick R, Farhadi A (2016) You only look once: unified, real-time object detection. In proceedings of the IEEE conference on computer vision and pattern recognition, pp 779-788 source
The paper states on p. 1 that it reframes object detection as a single regression problem, in contrast to prior approaches that repurpose classifiers. It specifies that YOLO predicts bounding-box coordinates and class probabilities, so the claim does not mean YOLO makes no class predictions. (LLM, conf=0.95)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“YOLOs are adopted in various applications majorly due to their faster inferences rather than considering detection accuracy.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Abstract
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption: Single-stage object detectors generally have faster inference than two-stage detectors.
- Probably not a substantial claim. 51
- It may state two separate claims as one. 30
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Feature Pyramid Network (FPN) [42] is one of the important two stage object detectors, tried to overcome few of the aforementioned challenges successfully by extracting features at multiple scales of an image in the object detection process.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Two stage object detection > R-CNN and successors
What the sources say
- Partly supportedcites [42] Lin TY, Dollár P, Girshick R, He K, Hariharan B, Belongie S (2017) Feature pyramid networks for object detection. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 2117–2125 sourceExperimental — accuracy not yet measured
The Abstract (p. 1) and Introduction (p. 2) support that FPN builds multi-scale features and improves object detection when used with Faster R-CNN. They describe FPN as a generic feature extractor used in detection systems, however, rather than as a two-stage detector in its own right. (LLM, conf=0.86)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“This version is faster and more accurate than all the previous versions.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v4)
What the sources say
- Partly supportedcites [5] Bochkovskiy A, Wang CY, Liao HY (2020) YOLOv4: optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934 sourceExperimental — accuracy not yet measured
The Introduction (p. 1) says YOLOv4 improves YOLOv3’s accuracy (AP) and speed (FPS) by 10% and 12%, respectively. The cited passages do not establish that it is faster and more accurate than every previous YOLO version. (LLM, conf=0.86)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“A novel OCR system is developed, named as Rosetta, wherein faster-RCNN is utilized for detecting the text characters from millions of Facebook images, however, fully convolutional CNN is employed for generating the lexicon free transcription of each word [6].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Two stage object detection > Applications of R-CNN and successors
What the sources say
- Partly supportedcites [6] Borisyuk F, Gordo A, Sivakumar V (2018) Rosetta: large scale system for text detection and recognition in images. In proceedings of the 24th ACM SIGKDD international conference on knowledge discovery data mining pp 71-79 sourceExperimental — accuracy not yet measured
The Introduction (p. 1) describes Rosetta as a Facebook OCR system that uses Faster-RCNN to detect individual words and a fully convolutional CNN to produce a lexicon-free transcription of each word. The claim is inaccurate in saying Faster-RCNN detects text characters; the paper assigns character recognition to the second stage. (LLM, conf=0.90)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“There are numerous other algorithms that have been introduced in recent past such as Single Shot Detector (SSD) [43], Deconvolution Single Shot Detector (DSSD) [16], RetinaNet [41], M2Det [86], RefineDet++ [85], are based on single stage object detection.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Introduction > Stages in object detection
What the sources say
- Supportedcites [43] Liu W, Anguelov D, Erhan D, Szegedy C, Reed S, Fu CY, Berg AC (2016) Ssd: single shot multibox detector. European Conf Comput Vis 2016:21–37. https://doi.org/10.48550/arXiv.1512.02325 sourceExperimental — accuracy not yet measured
The citation to this paper is attached specifically to SSD, not to the other named algorithms. The Conclusions (p. 15) identify SSD as a single-shot object detector, supporting its inclusion as a single-stage detection algorithm. (LLM, conf=0.94)
- Supportedcites [16] Fu CY, Liu W, Ranga A, Tyagi A, Berg AC (2017) Dssd: Deconvolutional single shot detector. arXiv preprint arXiv:1701.06659 source
The citation is attached to DSSD, not to the other separately cited algorithms. The Introduction (p. 1) identifies this paper’s proposed system as the deconvolutional single shot detector, built by augmenting SSD. (LLM, conf=0.94)
- Supportedcites [41] Lin TY, Goyal P, Girshick R, He K, Dollár P (2017) Focal loss for dense object detection. In: Proceedings of the IEEE international conference on computer vision, pp 2980–2988. https://doi.org/10.48550/arXiv.1708.02002 sourceExperimental — accuracy not yet measured
The Introduction (p. 1) presents RetinaNet as a one-stage object detector, and the RetinaNet Detector section (p. 4) describes its design for one-stage detection. Because this citation is attached to RetinaNet, it need not substantiate the other named algorithms. (LLM, conf=0.95)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“According to [65], the inception modules contain three different sizes of filters viz. 1 × 1, 3 × 3, and 5 × 5 for capturing different sized patterns.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > GoogLeNet
What the sources say
- Supportedcites [65] Szegedy C, Liu W, Jia Y, Sermanet P, Reed S, Anguelov D, Erhan D, Vanhoucke V, Rabinovich A (2015) Going deeper with convolutions. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 1–9. https://doi.org/10.48550/arXiv.1409.4842 source
Architectural Details (p. 4) says the Inception architecture uses 1 × 1, 3 × 3, and 5 × 5 filters. The same section (p. 5) explains that this design processes visual information at various scales, supporting the paraphrase about capturing different-sized patterns. (LLM, conf=0.94)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“It has resulted in a decrease in number of parameters, so as the model complexity [39].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > Network in network
What the sources say
- Supportedcites [39] Lin M, Chen Q, Yan S (2013) Network in network. arXiv preprint arXiv:1312.4400. https://doi.org/10.48550/arXiv.1312.4400 source
Section 3.2 (p. 4) says global average pooling replaces fully connected layers and has no parameters to optimize at that layer. Section 4.6 (p. 7) contrasts its fixed transformation with the dense, learned transformation of a fully connected layer, supporting the claim that this replacement reduces parameters and model complexity. (LLM, conf=0.87)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“The authors proposed this architecture in two flavors viz. VGG19 and VGG16, consisting of 19 and 16 layers in the deep neural networks respectively [64].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > VGG
What the sources say
- Supportedcites [64] Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409. https://doi.org/10.48550/arXiv.1409.1556 source
The paper calls its models the 16-layer Net-D and the 19-layer Net-E in “Generalisation of Very Deep Features” (p. 13). This supports the attributed layer counts for VGG16 and VGG19. (LLM, conf=0.93)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“The number of layers is increased to 34 which is almost doubled in comparison with VGG19 [64].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > ResNet
What the sources say
- Supportedcites [64] Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409. https://doi.org/10.48550/arXiv.1409.1556 source
Table 1 in §2.3 (p. 3) identifies VGG configuration E as having 19 weight layers, supporting the VGG19 comparison for which this citation is offered. The cited paper does not establish ResNet’s 34-layer count; that is not the VGG-specific point attributed to it here. (LLM, conf=0.88)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“The third version of YOLO tries to overcome aforementioned drawbacks and provides an efficient way for detecting objects with an improved performance while trained and evaluated on the COCO dataset [55].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v3)
What the sources say
- Supportedcites [55] Redmon J, Farhadi A (2018) Yolov3: an incremental improvement. arXiv preprint arXiv:1804.02767 source
Section 2.3 (p. 2) describes YOLOv3 experiments with COCO, and Section 3 (p. 3) evaluates its detection performance using COCO metrics. The Abstract (p. 1) says the new network was trained and is more accurate than the previous version while remaining fast. The improvement is metric-dependent: Section 3 notes that YOLOv3 trails RetinaNet on COCO’s average AP metric. (LLM, conf=0.87)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“They proposed smart factorization in the convolution operation, specifically, 3 × 3 filters are used throughout the entire architecture to introduce model consistency and to reduce the number of parameters.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > VGG
What the sources say
- Partly supportedcites [64] Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409. https://doi.org/10.48550/arXiv.1409.1556 sourceExperimental — accuracy not yet measured
The Discussion (p. 2) says the networks use small 3 × 3 receptive fields throughout and explains that stacking them uses fewer parameters than a single 7 × 7 layer. The Architecture section (p. 2) also notes a configuration with 1 × 1 filters, and the paper does not identify “model consistency” as a reason for the design. (LLM, conf=0.85)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Object detection algorithms in deep learning need larger datasets for computation, labor intensive approaches for annotations, and powerful computational resources for processing [45].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Introduction > Challenges in object detection
What the sources say
- Supportedcites [45] Liu L, Ouyang W, Wang X, Fieguth P, Chen J, Liu X, Pietikäinen M (2020) Deep learning for generic object detection: a survey. Int J Comput Vis 128(2):261–318. https://doi.org/10.48550/arXiv.1809.02165 sourceExperimental — accuracy not yet measured
Section 2.3 (p. 5) says deep detectors rely heavily on vast training data and high-capability GPUs, and that accurate annotations are labor-intensive to obtain. This supports all three parts of the attributed claim. (LLM, conf=0.94)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Additionally, SoftMax is used as an activation function in the last layer and (1 × 1) convolution filters are placed in a multilayer perceptron after convolutional layer resulting in dimensionality reduction [23].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > Network in network
What the sources say
- Partly supportedcites [23] Gu J, Wang Z, Kuen J, Ma L, Shahroudy A, Shuai B, Liu T, Wang X, Wang G, Cai J, Chen T (2018) Recent advances in convolutional neural networks. Pattern Recogn 77:354–377. https://doi.org/10.1016/j.patcog.2017.10.013 sourceExperimental — accuracy not yet measured
Section 3.1.4 (p. 5) says Network in Network places 1 × 1 convolutions after a traditional convolutional layer in its mlpconv layer and feeds the final pooled output into softmax. It does not say those convolutions reduce dimensionality; Section 3.1.5 (p. 5) assigns that role to 1 × 1 convolutions in the Inception module. (LLM, conf=0.90)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Real time object detection and tracking is implemented and benchmarking results are obtained using embedded hardware such as Jetson TX/AGX Xavier and Intel Neural Compute Stick wherein RCNN is employed on the embedded hardware for object detection [26].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Two stage object detection > Applications of R-CNN and successors
What the sources say
- Partly supportedcites [26] Hossain S, Lee DJ (2019) Deep learning-based real-time multiple-object detection and tracking from aerial imagery via a flying robot with GPU-based embedded devices. Sensors 19(15):3371 sourceExperimental — accuracy not yet measured
The Abstract and §7 describe embedded detection and tracking with benchmark results for Jetson devices and systems using a Neural Compute Stick. Section 2.2.1 names R-CNN among the detection algorithms for the Xavier system, and §5.1 reports that Faster R-CNN was accurate but had a very low performance rate. The cited passages do not establish that R-CNN ran on the Neural Compute Stick or that its detection achieved real-time performance; §5.6 reports stick-based FPS results without identifying R-CNN. (LLM, conf=0.83)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“It is 20 times deeper than AlexNet and 8 times deeper than VGG [23].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > ResNet
What the sources say
- Supportedcites [23] Gu J, Wang Z, Kuen J, Ma L, Shahroudy A, Shuai B, Liu T, Wang X, Wang G, Cai J, Chen T (2018) Recent advances in convolutional neural networks. Pattern Recogn 77:354–377. https://doi.org/10.1016/j.patcog.2017.10.013 source
The opening passage (p. 1) explicitly says ResNet is about 20 times deeper than AlexNet and 8 times deeper than VGGNet. The citing sentence conveys those same comparisons, though it omits “about.” (LLM, conf=0.94)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Supported
In the paper
“Fast R-CNN [20] achieves a significant improvement in the model training time and inference time. Moreover, it also observes an increase in the performance matrix of the object detection i.e., mAP.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Two stage object detection > R-CNN and successors
What the sources say
- Not checked — no full text of the sourcecites [20] Girshick R (2015) Fast r-cnn. In proceedings of the IEEE international conference on computer vision 2015, pp 1440–1448 source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“YOLO can be seen as giving a tough fight to not only two staged detectors but previous single staged detectors also in terms of accuracy and inference time.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Introduction > Stages in object detection
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
- Rests on an assumption: Single-stage object detectors generally have faster inference than two-stage detectors.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Moreover, with the advent of YOLO (You Only Look Once) and its architectural successors, the detection accuracy is improving significantly and sometime it is better than two stage detectors.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Abstract
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“In the experimentational setup for object detection on Pascal VOC 2007, Fast-RCNN and Faster-RCNN are 25 and 250 times faster respectively compared to traditional RCNN with almost similar mAP of around 66.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Two stage object detection > R-CNN and successors
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“A variant of YOLO with lesser model complexity known as Fast YOLO is proposed for faster detection of objects. It has 9 convolutional layers with comparatively lesser filters in those layers.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v1)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
In the paper
“Unfortunately, the network turned out to be an over parameterized network leading to large training error [25].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > ResNet
What the sources say
- Not checked — no full text of the sourcecites [25] He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. In proceedings of the IEEE conference on computer vision and pattern recognition, pp 770-778 source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Confidence score (cs) is computed for each bounding box per grid by multiplying pc with Intersection over Union (IoU) between the ground-truth and predicted-bounding-box.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“YOLO is not the first algorithm that uses Single Shot Detector (SSD) for object detection.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Introduction > Stages in object detection
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
None of the sources cited for this claim have been checked yet.
In the paper
“CSPDarknet53, a network with 29 convolution layers with 3 × 3 filters and approximately 27.6 million parameters, is chosen as backbone that outperformed the remaining architectures”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v4)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This architecture is based on Darkent-53 framework, a network that uses 3 × 3 and 1 × 1 convolutional filter along with some shortcut connections, significantly larger and powerful architecture with 53 convolutional layers.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v3)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“It is considered as one of the most common choices in production only because of its simple architectural design, low complexity, and easy implementation.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Introduction > Stages in object detection
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“the image/frame is divided into S × S grid cells, each grid cell predicts B bounding boxes along with their positions and dimensions, probability of an object in the underlying grid, and conditional class probabilities.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“In addition, ResNet152 has a total of 25.6 million parameters whereas ResNet110 has 1.7 million parameters [31].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > ResNet
What the sources say
- Not checked — no full text of the sourcecites [31] Khan A, Sohail A, Zahoora U, Qureshi AS (2020) A survey of the recent architectures of deep convolutional neural networks. Artif Intell Rev 53(8):5455–5516. https://doi.org/10.1007/s10462-018-9633-3 source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“It generates feature maps at three different scales. Specifically, it down-samples the input at 3 different scales by a factor of 32, 16, and 8.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v3)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Lighter the model, faster the inferences and vice versa, generating a trade-off between the efficacy and performance matrix of a deployed structured model on tensor processing unit (TPU).”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Bounding boxes are the approximations of the ground-truth. Generally, background pixels are also included during predictions, this affects the accuracy of the algorithm.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Introduction > Challenges in object detection
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“After exclusion of fully connected layers, YOLO (v2) can now operate on any dimension ranging from 320 × 320 to 608 × 608.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v2)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Here three layers of mlpconv are stacked followed by a Global Average Pooling (GAP) layer [12].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > Network in network
What the sources say
- Not matchedcites [12] Cook A (2017) Global average pooling layers for object localization. https://alexisbcook.github.io/2017/globalaverage-poolinglayers-for-object-localization/. Accessed 19 Aug 2019
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Not matched
In the paper
“This architecture is a deep network with 22 layers but has 12 times less parameters than AlexNet architecture, consisting of 4 million parameters approximately [31].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > GoogLeNet
What the sources say
- Not checked — no full text of the sourcecites [31] Khan A, Sohail A, Zahoora U, Qureshi AS (2020) A survey of the recent architectures of deep convolutional neural networks. Artif Intell Rev 53(8):5455–5516. https://doi.org/10.1007/s10462-018-9633-3 source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The architecture of YOLO is inspired by the GoogLeNet architecture.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v1)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“In addition, Path Aggregation Network (PAN) is preferred over Feature Pyramid Network (FPN) where PAN is a bottom-up path augmentation that uses Adaptive feature pooling which works similar to FPN.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v4)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“It records much better performance than any of the previous architectures on ImageNet dataset.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > GoogLeNet
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“In YOLO, we have 24 convolution layers followed by 2 fully connected layers as shown in Fig. 18.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v1)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“It uses darknet-19 framework consisting of 19 convolutional layers and 5 max pooling layers, as illustrated in Fig. 19.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v2)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“YOLOs are performing significantly better in comparison with their counterpart two stage object detectors in terms of detection accuracy and inference time.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Summarization and future research directions
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The second criteria for discarding the less relevant bounding boxes is known as non max suppression which is further based upon the IoU.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Subsequently, the concept of identity mapping is introduced in [11]. A skip or residual connection of length 2 is proposed as shown in Fig. 13.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > ResNet
What the sources say
- Not checked — no full text of the sourcecites [11] Choi H, Ryu S, Kim H (2018) Short-term load forecasting based on ResNet and LSTM. In IEEE international conference on communications, control, and computing Technologies for Smart Grids (SmartGridComm), pp 1-6 source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The improvised architecture is known as VGG, achieved a top-5 accuracy of 92.7% for an object detection task on the test dataset of ImageNet, a standard dataset containing around 14 million images belonging to 1 k classes.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > VGG
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Smaller filters result in fewer parameters and are capable of identifying smaller objects more effectively.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > VGG
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Though deep learning models outperform traditional machine learning approaches by a great margin, they demonstrate poor performance while evaluating on the datasets with fewer instances.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Introduction > Challenges in object detection
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Due to scaling inefficacy in deep neural networks, CNN is widely adopted to capture the spatial and contextual information with fewer parameters.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“A group of researchers from Google developed an application using computer vision, machine learning, and Google’s Knowledge Graph called as Google Lens [22] wherein a region proposal network (RPN) is utilized to detect the character level bounding boxes for text recognition.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Two stage object detection > Applications of R-CNN and successors
What the sources say
- Not matchedcites [22] Google Lens – Wikipedia (n.d.), https://en.wikipedia.org/wiki/Google_Lens. Accessed 06 Aug 2020
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Not matched
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“An approximate increase of 7% in recall is observed by the usage of convolution with anchor boxes, however, 0.3% reduction in mAP is also noticed.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v2)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“however, this version struggles in detection of smaller sized objects.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v2)
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
In the paper
“With increase in depth of network by series of convolutions, pooling layers and fully connected layers, it has resulted in 138 M parameters approximately [31].”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > VGG
What the sources say
- Not checked — no full text of the sourcecites [31] Khan A, Sohail A, Zahoora U, Qureshi AS (2020) A survey of the recent architectures of deep convolutional neural networks. Artif Intell Rev 53(8):5455–5516. https://doi.org/10.1007/s10462-018-9633-3 source
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Large localization error and lower recall in comparison with two stage object detectors may be considered as the two significant drawbacks of this version of YOLO.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v1)
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Authors introduced the problem of model degradation i.e. model accuracy increases with an increase in the model depth to a certain extent only.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > ResNet
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“As this version outperforms for smaller sized objects, however, suffers in producing accurate results for medium and large sized objects.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v3)
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“A grid cell predicts (B × 5 + n) values, where B is the number of bounding boxes per grid cell. The output tensor shape would be S × S×(B × 5 + n)”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs
hedgeddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“In order to overcome the problem of vanishing gradients, auxiliary loss is utilized by introducing an intermediate SoftMax classifiers during model training.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Convolutional neural networks and Pretrained models > GoogLeNet
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Moreover, YOLOs make comparatively fewer errors in the identification of the background”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Summarization and future research directions
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The combination of YOLO and Fast-RCNN outperform to the each of these standalone architecture on Pascal VOC dataset.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Summarization and future research directions
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Single stage object detectors are mainly designed for real time object detection; however, they are lagging in the performance metrics to the two stage object detectors.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Summarization and future research directions
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“YOLO-v4 is also an attempt to better deal with the trade-off between the accuracy and speed of the model by introducing several universal optimization features of CNNs.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Summarization and future research directions
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Generally, deep models take huge amount of time for model training due to the large dataset size and model complexity”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Summarization and future research directions
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“With the help of multi scale training, YOLO-v2 offers better detection on varying sized objects with an easy trade-off between the model performance and model inferences.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Summarization and future research directions
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“YOLOs are being rigorously employed in various real time object tracking such as self-driving cars.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Summarization and future research directions
boostedcorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Models trained for one particular task may not perform well on other similar tasks, resulting non-generalizability of the model for the data it has not seen before.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Summarization and future research directions
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“YOLO-v3 is better detector, fast, and accurate, necked by feature pyramid network.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Summarization and future research directions
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
The paper relies on this without support.
Most existing reviews and surveys cover two-stage object detection.
“Most of the reviews and surveys cover the two stage object detection algorithms.”
The authors state this as an assumption.
This is the first review covering single-stage object detection using YOLOs.
“To the best of our knowledge, this is the first review that covers the single stage object detection using specially YOLOs.”
The paper relies on this without support.
Single-stage object detectors generally have faster inference than two-stage detectors.
“However, the inference time of single stage detectors is better compared to its counterparts.”
The paper relies on this without support.
Most existing object detection techniques are restricted to classification for a limited number of categories.
“Most of the existing object detection techniques are restricted to classification for limited categories.”
The paper relies on this without support.
The scarcity of labelled data for object detection is the reason existing techniques are limited to few categories.
“This is certainly due to scarcity of labelled data for object detection task.”
The paper relies on this without support.
Random initialization is the biggest hurdle in predicting genuine offsets.
“Random initialization is considered as the biggest hurdles in predicting genuine offsets.”
The paper relies on this without support.
Better shaped anchor boxes provide a quick start and improve model training and performance.
“Better shaped anchor boxes may provide quick start and offers improved model training and performance.”
The paper relies on this without support.
The threshold for discarding boxes is usually 0.5 in most object detection algorithms.
“Usually, this threshold is taken as 0.5 in most of the object detection algorithms”
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“Machine learning models, particularly those based on deep learning architectures like YOLO, have demonstrated significant advancements in object detection capabilities [93].”
From A Review of Passenger Counting in Public Transport Concepts with Solution Proposal Based on Image Processing and Machine Learning (2024) · cites this paper as [93]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v3)
“Conclusively different feature maps viz. 52 × 52, 13 × 13, and 26 × 26 are utilized for detecting large, smaller, and medium-sized objects respectively.”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Architectural design of YOLOs > YOLO (v3)
“Finally, the 106th layer is involved in the detection using a 52 × 52 feature map after applying a stride of 8. Following the same process, adding few convolutions to the 91st layer and combining with the 36th layer using (1 × 1) kernel, the down-sampled feature maps are concatenated to Up-sampled feature maps at different places to extract fine-grained features for detecting smaller objects of various dimensions.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Two stage object detection > R-CNN and successors
“Fast R-CNN [20] achieves a significant improvement in the model training time and inference time. Moreover, it also observes an increase in the performance matrix of the object detection i.e., mAP. Single stage object detection is realized with the help of multi-task loss function”
§ Object detection using YOLO: challenges, architectural successors, datasets and applications > Summarization and future research directions
“In this paper, we explored two stage object detectors viz. RCNN, Fast-RCNN, and Faster-RCNN along with their important applications.”
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 3×
“20 Skip connections in the ResNet Module [25] aforementioned drawbacks and provides an efficient way for detecting objects with an improved performance while trained and evaluated on the COCO dataset [55].”
BackgroundVery Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 2×
“VGG19 and VGG16, consisting of 19 and 16 layers in the deep neural networks respectively [64].”
BackgroundResultYou Only Look Once: Unified, Real-Time Object Detection
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.91 · Cited by 53,803 · cited 2×
“18 YOLO architecture for object detection and localization [56] predicts there ’ s an object.”
BackgroundRich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation
Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.81 · Cited by 32,684 · cited 1×
“R-CNN is a region based convolutional neural network object detection algorithm proposed by Ross Girshick [21].”
MethodFeature Pyramid Networks for Object Detection
Lin, T.-Y., Dollar, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.106 · Cited by 30,939 · cited 6×
“It incorporates heuristics like residual blocks, skip connections, and up-sampling similar to FPN.”
BackgroundInfluentialGirshick, R. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.169 · Cited by 28,105
SSD: Single Shot MultiBox Detector
Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.-Y., & Berg, A. C. (2015) · arXiv · doi:10.48550/arxiv.1512.02325 · Cited by 22,443 · cited 1×
“There are numerous other algorithms that have been introduced in recent past such as Single Shot Detector (SSD) [43], Deconvolution Single Shot Detector (DSSD) [16], RetinaNet [41], M2Det [86], RefineDet++ [85], are based on single stage object detection.”
MethodSSD: Single Shot MultiBox Detector
Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.-Y., & Berg, A. C. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46448-0_2 · Cited by 22,128
YOLOv3: An Incremental Improvement
Redmon, J., & Farhadi, A. (2018) · arXiv · doi:10.48550/arxiv.1804.02767 · Cited by 17,141
Robust Real-Time Face Detection
Viola, P., & Jones, M. J. (2004) · International Journal of Computer Vision · doi:10.1023/b:visi.0000013087.49260.fb · Cited by 14,097 · cited 1×
“Well known machine learning approaches for object detection include Viola-Jones object detection framework [70] and Histogram of Oriented Gradients (HOG).”
Method
Research works printed but not linked: 4
- Morris T (2004) Computer Vision and Image Processing, Palgrave Macmillan Ltd, 1st edition, pp 1–320
- Ren S, He K, Girshick R, Sun J (2015) Faster r-cnn: towards real-time object detection with region proposal networks. Adv Neural Inform Process Syst:91–99
- Tsang S-H (2018) Review: Inception-v4 - Evolved From GoogLeNet, Merged with ResNet Idea (Image Classification), towards data science
- Zaytar MA, El Amrani C (2016) Sequence to sequence weather forecasting with long short-term memory recurrent neural networks. Int J Comput Appl 143(11):7–11
Cited by
Sorted by citations
Items 1,201–1,300 of 1,537
Zhang, X., Xue, J., & Zhao, S. (2026) · Symmetry · doi:10.3390/sym18040561
Review on Cervical Spine Fracture Detection Using Deep Neural Networks
Puni, S. K., Embadi, S., Karagalla, S., et al. (2025) · Lecture notes in networks and systems · doi:10.1007/978-3-031-78922-9_34
Temporal integrated CNN for robust real time video based traffic sign recognition on indian roads
Kandasamy, K., Natarajan, Y., Preethaa, K. R. S., et al. (2026) · Neural Computing and Applications · doi:10.1007/s00521-026-12121-z
Sabthami, J., Ramani, B., & Vijayalakshmi, P. (2026) · The Journal of Supercomputing · doi:10.1007/s11227-026-08468-8
Harish, S., Nawar, S., Chaitra, B. N., Amin, R., & Vijayan, A. (2025) · Lecture notes in networks and systems · doi:10.1007/978-981-96-4008-9_44
Zhai, G., & Jiang, Y. (2024) · International Conference on New Trends in Computational Intelligence (NTCI) · doi:10.1109/ntci64025.2024.10776408
Automatic Detection of Serve Actions in Tennis Videos
Choroś, K. (2024) · Communications in computer and information science · doi:10.1007/978-981-97-5934-7_4
Accurate Positioning and Motion Detection by Neural Metalens-Integrated Camera
Hao, C., Yuan, Z., Chen, J., et al. (2025) · doi:10.1109/wcsp68525.2025.1010241
Cargo positioning and sorting model based on deep learning and sampling evaluation
Xue, B., & Zhu, Z. (2026) · Systems and Soft Computing · doi:10.1016/j.sasc.2026.200441
Qin, Y., Chen, L., Mao, Y., & Zhou, M. (2024) · Multimedia Tools and Applications · doi:10.1007/s11042-024-19966-3
A Lightweight System for Strawberry Disease Detection and Segment Based on YOLOv8
Cao, Z., Wang, Y., Jiang, Z., et al. (2024) · International Conference on Electronic Information Engineering and Computer Science (EIECS) · doi:10.1109/eiecs63941.2024.10800290
“YOLO is a well-established object detection model in deep learning [7,8].”
Vision-cart: IoT - Enabled Smart Glasses for an Enhanced Shopping Experience
Poovizhi, P., Vijaya Kumar, T., Rajasekaran, P., et al. (2025) · doi:10.1109/icima64861.2025.11074045
Bayineni, V. K., Gouda, G. M., Tekur, S. H., et al. (2026) · ISSS Journal of Micro and Smart Systems · doi:10.1007/s41683-026-00150-8
Chen, H., Qiu, F., Kulsum, U., Zhang, C., Bohnett, E., Lyu, H., Tian, S., Lamichhane, B. R., Stow, D. A., Lewison, R. L., et al. (2026) · European Journal of Wildlife Research · doi:10.1007/s10344-026-02156-x · Cited by 1
“On the other hand, one-stage detectors, such as the YOLO series, are designed for real-time detection, providing rapid processing speed and high efficiency while maintaining strong detection accuracy (Diwan et al. 2023 ; Ferrante et al. 2024).”
Computer Localization Algorithms in Visual Recognition Software
Yuan, Y. (2024) · Procedia Computer Science · doi:10.1016/j.procs.2024.10.095
Angelino, A., D'Amato, D., Galante, G. A., Landolfi, E., Massa, A. A., & Troiano, A. (2025) · IEEE International Conference on Quantum Artificial Intelligence (QAI) · doi:10.1109/qai63978.2025.00074
Ge, H., & Hu, L. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-01018-4
Straehle, J., Galanis, C., Grünewald, L., Balta, E., Deller, T. D., Häussler, U., Mizaikoff, B., Beck, J., & Vlachos, A. (2026) · Frontiers in Synaptic Neuroscience · doi:10.3389/fnsyn.2026.1771781
“Expert-labeled datasets derived from TEM and FIB-SEM images are used to train and refine automated segmentation pipelines based on deep learning approaches, for example, synaptic vesicle detection using YOLO-based object detection (Diwan et al., 2023).”
Real-Time Automated Basketball Highlight Generation Using YOLO-Based Object Tracking
Subramanian, R. R., Sreevastha, T. V., Thisyanth, P., et al. (2025) · doi:10.1109/iccrtee64519.2025.11052994
Enhancing Used Car Price Predictions with Machine Learning-Based Damage Detection
Kulkarni, R. V., Thopate, K., Khatib, F., Dixit, A., Ingle, A., & Kanathia, A. (2024) · IEEE Global Conference for Advancement in Technology (GCAT) · doi:10.1109/gcat62922.2024.10923877
“Some of the multi-stage detectors are Residual Network 50 (ResNet50) [15], Region-based Convolutional Neural Network (R-CNN) [17], Faster R-CNN [18], and Mask R-CNN [19].”
Peng, Y., Yang, L., Yang, H., Wu, T. H., Li, X., & Zhang, J. (2025) · Food Control · doi:10.1016/j.foodcont.2025.111580
AI‐ and Image‐Based Analysis of Emulsification Processes: Opportunities and Challenges
Burke, I., Derkum, S., Olusanya, T. O. O., et al. (2025) · Chemie Ingenieur Technik · doi:10.1002/cite.70018
Research on an improved RT-DETR-based model for rice disease detection
Zhang, Y., Shen, C., & Xiong, Y. (2026) · PLoS ONE · doi:10.1371/journal.pone.0349237
Indian Sign Language Recognition using YoloV5 on Custom Dataset
Gautam, A., Aishwaryam, A., Kaur, S., et al. (2026) · doi:10.1109/icsped65817.2026.11448389
“An essential issue in computer vision is object detection, which has evolved over the years as an outcome of continuous development of more precise and efficient algorithms [5] [6].”
Real-Time Traffic Light Optimization Using Yolov9 and Length-Based Metrics
Korade, N. B., Salunke, M. B., Bhosle, A. A., et al. (2025) · International Journal of Electrical and Electronics Research · doi:10.37391/ijeer.130212
Early-Stage Fire Recognition and Localization Based on YOLOv8 and Binocular Vision Fusion
Gan, Y., Xiong, L., Xu, W., et al. (2026) · doi:10.1109/rpic68328.2026.11518089
“With the evolution of its architecture, newer versions of YOLO have demonstrated substantial improvements in detection accuracy, sometimes outperforming two-stage detectors [3].”
Slimane, J. B. (2024) · International Journal on Information Technologies and Security · doi:10.59035/zbte3810
Deteksi Objek Menggunakan Metode YOLO dan Implementasinya pada Robot Bawah Air
Husnan, H., Fatichah, C., & Dikairono, R. (2023) · Jurnal Teknik ITS · doi:10.12962/j23373539.v12i3.122326
Toward Causal and Evidential Open-Set Temporal Action Detection
Wang, Z., Zhao, R.-W., Feng, R., & Jin, C. (2025) · IEEE Access · doi:10.1109/access.2025.3553717
Al-Azzam, Q. H. A., Balachandran, W., & Hunaiti, Z. (2026) · AI in Medicine · doi:10.3390/aimed1030024 · Cited by 1
Oda, Y. S. (2023) · Digital Library of Theses and Dissertations (Universidade de São Paulo) · doi:10.11606/t.76.2023.tde-22022024-084207
Blockly–ROS Platform for Libras-Based Robotic Navigation with Emphasis on Laterality Training
Holtz, A. G., Pinel, G. S. F., & Brandao, A. S. (2025) · doi:10.1109/icar65334.2025.11338740
“Models like YOLOv8n can perform detection and classification at around 30 FPS on embedded GPUs, maintaining a mean average precision (mAP) above 90% for static gesture datasets [13], [14], making them suitable for educational applications that require fast and accurate responses [15], [16].”
Shang, Z., Zhao, M., Yong, W., et al. (2026) · doi:10.1109/sgai69026.2026.11566025
Automated Classification of Cocoa Bean Fermentation Levels Using Computer Vision
Suarez, J., Espinosa, J., Contreras, K., et al. (2025) · doi:10.1109/stsiva66383.2025.11156348
“The YOLO architecture is based on dividing the input image into a grid, and each cell of this grid is responsible for predicting one or more bounding boxes and the probabilities of the classes associated with the detected objects [14].”
Xu, P., Fan, R., Feng, Y., et al. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6409037
Tassan-Din, E., Gunessee, A., Vong, P., Marqué, C., Dolla, L., Picar, A. M., & Monstein, C. (2026) · Journal of Space Weather and Space Climate · doi:10.1051/swsc/2026014
Vision-AI-driven dilemma zone modelling for driver behaviour prediction
Psalta, A., Kopsacheilis, A., Tsironis, V., Politis, I., & Karantzalos, K. G. (2026) · European Transport Research Review · doi:10.1186/s12544-026-00836-y
“Recent advancements in object detection [7, 75], multiple object tracking [11, 51, 54], trajectory prediction [50, 61] and object re-identification [55, 73] have enabled the development of robust, efficient and fully automated computer vision pipelines, which are typically based on video data from…”
Ye, C., Wu, M., Yang, Y., et al. (2024) · doi:10.1109/icdcot61034.2024.10515672
Enhance YOLOV5 Model to Detect the Ocean Ships Using Real-Time Hyperspectral Satellite Image
Dhivya, K., Venkatesan, M., & Prabhavathy, P. (2025) · Lecture notes in electrical engineering · doi:10.1007/978-981-96-1587-2_10
Silveira, I. F., Cardoso, A., de Siqueira, A. G., Martins, V. F., Eliseo, M. A., Neto, J. S. D. O., Venkatakrishnan, R., Grill, B. M., Varshavsky, A., & Lok, B. C. L. (2026) · Lecture notes in educational technology · doi:10.1007/978-981-95-7580-0_13
Zhu, Z., Ke, F., Rahman, A., & Wang, Y. (2025) · IEEE Access · doi:10.1109/access.2025.3617226
MBF-DETR: a lightweight real-time detector with multi-backbone hybrid attention for ship detection
Fan, X., Xing, B., Wang, X., et al. (2026) · Journal of Real-Time Image Processing · doi:10.1007/s11554-026-01896-y
“Deep learning significantly advanced object detection through automatic feature learning [20, 21].”
A Study on Fine-Grained Motorcycle Classification for Intelligent Transportation Systems
Lima, G. E., Santos, E., Nascimento, E., Jr., et al. (2025) · doi:10.5753/sibgrapi.est.2025.38290
Automated Blood Group Detection Using YOLOv11x: A Deep Learning Approach
Mansoor, Z. K., Ali, M. H., & Al‐Askari, M. A. (2026) · Lecture notes in networks and systems · doi:10.1007/978-3-032-13921-4_10
Yang, W., Wang, Q., Li, E., Hu, Z., Hu, Y., Peng, W. J., & Sun, J. (2026) · Applied Sciences · doi:10.3390/app16189342
Doroshenko, A., Глод, С. І., & Kynash, Y. (2026) · Technologies · doi:10.3390/technologies14100604
Object Detection In Video Frames with TimestampingPreprint
Raafee, M. I., Lasrado, D., Naik, P. P., et al. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5065756
Carnaghi, M., Mostafa, K., Hany, M., et al. (2026) · Acta Tropica · doi:10.1016/j.actatropica.2026.108120
Fine-Tuning YOLOv8 for Enhanced Accuracy in Marine Environments
Mishra, S., Kaushik, T., Sonthaliya, U., et al. (2024) · IEEE Region 10 Conference (TENCON) · doi:10.1109/tencon61640.2024.10902884
Shafiee-Nattaj, M., Ghazipoor, K., & Hosseini‐Sarvari, M. (2026) · Biomass and Bioenergy · doi:10.1016/j.biombioe.2026.109802
He, W. P., & Li, Q. (2025) · International Conference on Computer Vision, Application, and Algorithm (CVAA) · doi:10.1117/12.3055958
“Single-stage algorithms, such as You Only Look Once (YOLO) [3] and Single Shot MultiBox Detector (SSD) [4] , predict object class and location in a single step, enabling fast detection.”
Fanariotis, A., Orphanoudakis, T. G., & Fotopoulos, V. E. (2025) · doi:10.1109/dsp65409.2025.11074952
“Renesas includes an example of such algorithm and utilizes the YOLO (You Only Look Once) model [8] for facial recognition on the EK-RA8D1 evaluation kit.”
Terengganu, U. M., Multajam, R., Ayob, A. F. M., Terengganu, U. M., Jamaludin, S. N. F., Terengganu, U. M., Sanjaya, W. S. M., Djati, U. I. N. S. G., Sambas, A., Tasikmalaya, U. S. Z. A. U. M., et al. (2025) · Mathematical Modeling and Computing · doi:10.23939/mmc2025.02.682 · Cited by 1
Cascade Deep Learning for High-Accuracy National License Plate Identification
Tabarnoust, E., & Zaz, Y. (2026) · Lecture notes in networks and systems · doi:10.1007/978-3-032-28573-7_2
Survey of Techniques and Applications in Object Recognition
Adnan, M. (2024) · International Journal for Research in Applied Science and Engineering Technology · doi:10.22214/ijraset.2024.62130
Detection of Candidate Wuyi Rock-Tea Shoots in Field RGB Images Using YOLO11s-CSNG
Qin, H., Yang, X., Yuan, J., et al. (2026) · Horticulturae · doi:10.3390/horticulturae12081024
“Reviews of YOLO development identify feature fusion, detection-head design, loss functions, and efficiency as major directions of architectural evolution [38].”
Yousefi, M., Dalvand, M. J., Khomami, N. A., et al. (2024) · International Conference on Control, Instrumentation, and Automation · doi:10.1109/iccia65044.2024.10768182
Khan, M. N., Armoush, R., Esfahani, A., Anisi, M. H., & Goudarzi, S. (2026) · IEEE Access · doi:10.1109/access.2026.3718358 · Cited by 1
MCS-DET: A Multiscale Object Detection Model for UAV Infrared Images Based on Edge EnhancementPreprint
Wu, H., Ye, S., Wang, Q., Ullah, F., & yonghong, Z. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5871352
Log What Matters: Safety-Aware Adaptive Logging for CAVs in Dynamic Conditions
Polin, J. A., Luo, Y., Sumaiya, F., Dong, Z., & Lu, S. (2026) · IEEE International Conference on Mobility, Operations, Services and Technologies (MOST) · doi:10.1109/most69733.2026.00027
“For camera-based detection, one-stage models such as SSD and YOLO deliver real-time performance [67], [68], while two-stage detectors such as Faster R-CNN improve precision through region-based processing [69]–[72]; transformer-based detectors further enhance context reasoning and global feature…”
Yang, R., Zhao, H., Kong, Y., et al. (2026) · Eng—Advances in Engineering · doi:10.3390/eng7070315
Jenrette, J., Hart, D. R., Chang, J., et al. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7304013
Holst, F., Gülsoylu, E., & Frintrop, S. (2026) · IEEE Journal of Oceanic Engineering · doi:10.1109/joe.2026.3695330
“YOLO models are widely recognized for their real-time object detection capabilities, thanks to their efficient single-stage detection approach that balances speed and accuracy [36].”
Neural Network-Based Modeling and Optimization of Vibration Damper Systems
Cheng, D., Luo, F., & Wang, C. (2025) · International Conference on Artificial Intelligence and Electromechanical Automation (AIEA) · doi:10.1109/aiea66061.2025.11160562
“These methods have achieved remarkable success in the fields of image recognition[5], natural language processing[6], autonomous driving[7] and object detection[8].”
Chauhan, A. S., & Nehra, V. (2025) · doi:10.1109/iemecon69302.2025.11365634
Batyuk, A. Y., & Опалінський, І. Ю. (2025) · Науковий вісник НЛТУ України · doi:10.36930/40350115
Object Detection in Video Frames with Timestamping
Suketha, Naik, P. P., Lasrado, D., Gowda, S., & Raafee, M. I. (2025) · Lecture notes in networks and systems · doi:10.1007/978-3-032-05545-3_17
Implementing YoloV8 on Zynq UltraScale + MPSoC FPGA Boards
Ballout, H., Berta, R., Srour, O., et al. (2026) · Lecture notes in electrical engineering · doi:10.1007/978-3-032-17305-8_31
GFE-Det: Global Feature Enhanced Method for Low-Altitude Target Detection
Chen, M., Miao, Z., & Liu, K. (2025) · Lecture notes in electrical engineering · doi:10.1007/978-981-96-2232-0_26
YOLOv11 based algorithm for abandoned luggage detection with dynamic radius estimation
Vrsalovic, I., Ivašić-Kos, M., Pobar, M., et al. (2025) · doi:10.1109/picom68402.2025.00023
VisuAid-Guiding Every Step, Illuminating Every Word
Shruthi, G. R., Tokala, S., & Chaithra, S. (2026) · Lecture notes in electrical engineering · doi:10.1007/978-981-95-5831-5_19
Retrofitting a Legacy Cutlery Washing Machine Using Computer Vision
Fwa, H. L. (2024) · Communications in computer and information science · doi:10.1007/978-3-031-70259-4_23
Mine Exogenous Fire Detection Algorithm Based on Improved YOLOv9
Zhan, X., Yao, R., Qi, Y., et al. (2026) · Processes · doi:10.3390/pr14010169
A YOLOv5-Based Printed Circuit Board Defect Detection System
Lin, J., & Chong, L. W. (2025) · IEEE International Conference on Consumer Electronics - Taiwan (ICCE-Taiwan) · doi:10.1109/icce-taiwan66881.2025.11207944
“YOLOv5, developed in Python, integrates modules such as Focus, CSPNet, SPP, and PANet, enhancing feature extraction efficiency and inference accuracy [4]-[6].”
Singh, A., & Dass, S. (2025) · International Journal of Intelligent Transportation Systems Research · doi:10.1007/s13177-025-00508-6
“In contrast to traditional traffic management plans based on stationary sensors, fixed heuristics, or rule-based reasoning [ 2–4], this paper presents a dynamic, deep learning-based framework with real-time adaptation to changing road interactions. patterns.”
Maruf, M. O. F., Islam, M. R., Rafi, M. J. A., Shake, M. A. S., & Islam, M. T. (2024) · International Conference on Electrical Engineering and Information Communication Technology · doi:10.1109/iceeict62016.2024.10534325
Wang, X., He, Z., Jiang, Y., et al. (2026) · Ecological Informatics · doi:10.1016/j.ecoinf.2026.103762
Face mask detection model using deep learning on edge computing
Chan, Y.-W., Chiao, H.-T., Oztepe, C., et al. (2026) · The Journal of Supercomputing · doi:10.1007/s11227-026-08594-3
Improve the Identification Accuracy of Parking Stickers with Persistent Homology Features
Zhang, X., Zhong, W., & Gu, F. (2025) · doi:10.1109/icicr65456.2025.00148
Research on Intelligent Grading of Appearance Quality of Red Fuji Apple Based on Improved YOLOv11
Pan, Y., Zhao, L., Hou, Z., et al. (2026) · International Conference on Computer Information Science and Application Technology (CISAT) · doi:10.1109/cisat71030.2026.11682290
“…follows the typical one-stage detection paradigm and is organized around three functional components: a backbone for hierarchical feature extraction, a neck for cross-scale feature aggregation, and a detection head for category prediction, bounding-box regression, and confidence estimation[14].”
Shaik, A., Tiwari, A., Jain, P., Kumar, S., Trujillo, C., & Banoth, E. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.110469
Breast Cancer Classification: A Semantic Comparison of YOLOv8 and YOLOv11
Singh, P., & Kumar, S. (2025) · Lecture notes in networks and systems · doi:10.1007/978-981-96-7134-2_26
Rahimighazvini, H., & Zargarzadeh, H. (2025) · International Journal of Research Publication and Reviews · doi:10.55248/gengpi.6.0325.1224
Tomczyk, N., Prokop, A., Borowiec, A., et al. (2026) · ce/papers · doi:10.1002/cepa.70753
“These values are normalized to the range ⟨ 0,1 ⟩ with respect to the image dimensions and reverse coordinate system, ensuring that the detection process is independent of absolute image resolution [8, 9].”
Drone-Assisted Pallet Tracking and Counting System Using YOLOv8
Akar, F., Kaya, S., Bayılmış, C., et al. (2025) · Innovations in Intelligent Systems and Applications Conference (ASYU) · doi:10.1109/asyu67174.2025.11208287
Safety Equipment Detection at Construction Site Using YOLOv8 and Varied Color Spaces
Pati, N., Sharma, A., Jena, J. J., Gourisaria, M. K., Parui, S., & Bandyopadhyay, A. (2026) · Lecture notes in networks and systems · doi:10.1007/978-981-95-6013-4_1
Improved YOLOv8 for early detection of Alzheimer’s disease from magnetic resonance imaging
Nguchu, B. A., Han, J., Mwighusa, D. N., Li, L., Su, K., & Shaw, P. M. (2026) · Frontiers in Aging Neuroscience · doi:10.3389/fnagi.2026.1771536
“This will not only speed up the learning process of the model but also reduce computation costs while improving detection performance (Diwan et al., 2022; Sirisha et al., 2023).”
Jing, D., Mi, L., & Gao, M. (2026) · International Conference on Integrated Intelligence and Communication Systems (ICIICS) · doi:10.1109/iciics67880.2026.11483515
Enhancing Seabass Detection in Aquaculture: A Step Toward Automated Behavioral Analysis Using AI
Nallarasan, N. K., Sharifonnasabi, F., Long, N. T., & Makhdoom, I. (2025) · Lecture notes in networks and systems · doi:10.1007/978-981-96-1758-6_10
Using YOLO Few_Shot Learning for Labeling Inflammatory Bowel Diseases
Fiaidhi, J., & Mohammed, S. (2025) · International Conference on Future Telecommunications and Artificial Intelligence (IC-FTAI) · doi:10.1109/ic-ftai67960.2025.11384536
“…of YOLO can be used for this purpose (e.g. Average Precision (AP) and mean Average Precision (mAP) ) depends on the IoU score ( Intersection over Union ) for the area between the predicted bounding boxes produced by the YOLO model and the ground-truth bounding boxes provided by the expert [13].”
Review on Development of E-Waste Management System
Burange, R. A. (2024) · International Journal for Research in Applied Science and Engineering Technology · doi:10.22214/ijraset.2024.58542
Forster, A., Lucheroni, C., & Gürtler, S. (2024) · International Symposium on Digital Forensics and Security (ISDFS) · doi:10.1109/isdfs60797.2024.10527261 · Cited by 1
Ghosh, G. K., Chatterjee, B., Kayal, S., et al. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-17181-8_28
Deep Learning–Based Corn Yield Component Estimation Under Different Nitrogen and Irrigation Rates
Ghimire, B., Lacerda, L. N., Bourlai, T., et al. (2026) · AgriEngineering · doi:10.3390/agriengineering8040146
“However, when comparing inference speed, YOLO was still approximately 300 times faster [82].”
Research on a UAV photogrammetry-based construction site waste detection & monitoring system
Ahamed, M., Gu, H., Li, F., et al. (2026) · Measurement Science and Technology · doi:10.1088/1361-6501/ae3ac8
“The most recent iteration of YOLO, YOLOV8, has shown improved capacity to detect and identify objects in challenging conditions, including construction waste, due to its more innovative architecture and attention mechanisms (Diwan et al 2022).”
Improving the unmanned aerial vehicle (UAV) image object detection algorithm based on YOLOv8s
Du, Y., & Yu, W. (2024) · International Conference on Intelligent Informatics and Biomedical Sciences (ICIIBMS) · doi:10.1109/iciibms62405.2024.10792734
YOLO Meets FedAVG: A Privacy-Preserving Approach to Autonomous Vehicles Object Detection
Alahdal, N. M., Abukhodair, F., Meftah, L. H., & Cherif, A. (2025) · International Conference on Advanced Innovations in Smart Cities (ICAISC) · doi:10.1109/icaisc64594.2025.10959368
“However, several challenges remain, particularly concerning data privacy, model scalability, and the need for consistent performance across diverse environments [4].”
Metadata
- Authors
- Diwan, T., Anirudh, G., & Tembhurne, J. V.
- Affiliations
- Indian Institute of Information Technology, Nagpur (IN)
- Central University of Rajasthan (IN)
- Corresponding author
- Diwan, Tausif
- Venue
- Multimedia Tools and Applications
- Published
- 2022-08-08
- Volume
- 82
- Issue
- 6
- Pages
- 9243-9275
- ISSN
- 1380-7501 (ISSN-L)
- 1573-7721
- PMC
- PMC9358372
- OpenAlex
- W4291819654
- Cited by
- 1,549
- References
- 73
- Citations by year
- 2026: 308 · 2025: 542 · 2024: 468 · 2023: 218 · 2022: 13
- FWCI
- 101.55
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Licence
- Locations
- Multimedia Tools and Applications (publishedVersion)
- PubMed Central (publishedVersion)
- PubMed (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Visual Attention and Saliency Detection
- Keywords
- object detection, YOLO, Fast R-CNN, inference latency, detection accuracy, real-time object detection
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-09-28, under version 3 of the claim definition (prompt extract-v3).
A second model labelled each claim on 2026-09-28 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
