You Only Look Once: Unified, Real-Time Object Detection
Open accessThe DOI matches its Crossref record; 53 of the 53 references it cites were found in a registry or an index; 8 of the 25 sources its claims cite were checked for support.
Abstract
We present YOLO, a new approach to object detection. Prior work on object detection repurposes classifiers to perform detection. Instead, we frame object detection as a regression problem to spatially separated bounding boxes and associated class probabilities. A single neural network predicts bounding boxes and class probabilities directly from full images in one evaluation. Since the whole detection pipeline is a single network, it can be optimized end-to-end directly on detection performance. Our unified architecture is extremely fast. Our base YOLO model processes images in real-time at 45 frames per second. A smaller version of the network, Fast YOLO, processes an astounding 155 frames per second while still achieving double the mAP of other real-time detectors. Compared to state-of-the-art detection systems, YOLO makes more localization errors but is less likely to predict false positives on background. Finally, YOLO learns very general representations of objects. It outperforms other detection methods, including DPM and R-CNN, when generalizing from natural images to other domains like artwork.
Overview
CoChat Summary
TL;DR
YOLO reframes object detection as a single regression problem, using one convolutional network to predict bounding boxes and class probabilities directly from full images. It achieves real-time detection speeds while remaining competitive with, and complementary to, slower region-based detectors.
Key findings
- YOLO runs at 45 FPS on a Titan X GPU, and Fast YOLO runs at 155 FPS; Fast YOLO reaches 52.7% mAP on PASCAL VOC 2007, more than double prior real-time detectors, while YOLO reaches 63.4% mAP.
- On PASCAL VOC 2012, YOLO scores 57.9% mAP; combining Fast R-CNN with YOLO raises Fast R-CNN from 71.8% to 75.0% mAP on VOC 2007, a 3.2% gain.
- YOLO makes fewer background false positives than Fast R-CNN: Fast R-CNN has 13.6% background false positives among top detections and is almost 3× more likely to predict background detections than YOLO.
- YOLO generalizes better to artwork: on Picasso, YOLO gets 53.3 AP versus R-CNN’s 10.4 AP and DPM’s 37.8 AP; on People-Art, YOLO gets 45 AP versus R-CNN’s 26 AP and DPM’s 32 AP.
Methods
- Detection is modeled as a single CNN regression: the image is divided into an S×S grid; for PASCAL VOC, S=7, B=2, C=20, producing a 7×7×30 tensor of bounding boxes, confidence scores, and class probabilities.
- The base network has 24 convolutional layers followed by 2 fully connected layers; Fast YOLO uses 9 convolutional layers. Convolutional layers are pretrained on ImageNet at 224×224, then detection training uses 448×448 input.
- Training uses PASCAL VOC 2007 and 2012 for about 135 epochs, batch size 64, momentum 0.9, decay 0.0005, sum-squared error with λ_coord=5 and λ_noobj=0.5, dropout, and data augmentation.
Limitations
- The paper states YOLO imposes strong spatial constraints because each grid cell predicts only two boxes and one class, limiting nearby objects and struggling with small objects in groups such as flocks of birds.
- The paper states YOLO struggles to generalize to objects in new or unusual aspect ratios or configurations, uses relatively coarse features due to downsampling, and its loss treats small and large box errors equally; its main error source is incorrect localization.
- The paper states YOLO still lags state-of-the-art detection accuracy and struggles to precisely localize some objects, especially small ones.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures





Verification
Claims the paper makes
6 Valid · 8 Not valid · 15 Not checked · 92 Own results and interpretations
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“While it is much faster than R-CNN, it still falls short of real-time and takes a significant accuracy hit from not having good proposals.”
What the sources say
- Could not verifycites [20] [20] K. Lenc and A. Vedaldi. R-cnn minus r. arXiv preprint arXiv:1506.06981 , 2015. 5, 6 - sourceExperimental — accuracy not yet measured
The ConvNet-based Object Detection section identifies R-CNN minus R as a method that uses predefined regions instead of region proposals. It gives no runtime measurement or assessment of whether the method achieves real-time detection; searches for real-time performance and speed did not establish that claim. (LLM, conf=0.76)
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: passes — The paper gives a reason why it holds.
- Could not verify
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“To avoid overfitting we use dropout and extensive data augmentation.”
What the sources say
- Partly supportedcites [18] [18] G. E. Hinton, N. Srivastava, A. Krizhevsky, I. Sutskever, and R. R. Salakhutdinov. Improving neural networks by preventing co-adaptation of feature detectors. arXiv preprint arXiv:1207.0580 , 2012. 4 - sourceExperimental — accuracy not yet measured
The paper states that dropout reduces overfitting (p. 1) and describes using dropout with data augmentation for its ImageNet model (H Models for ImageNet, p. 18). It does not establish that YOLO uses either technique or describe YOLO’s augmentation as extensive. (LLM, conf=0.88)
plainstates a cause- 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
“However, only 30Hz DPM [31] actually runs in real-time.”
What the sources say
- Supportedcites [31] [31] M. A. Sadeghi and D. Forsyth. 30hz object detection with dpm v5. In Computer Vision-ECCV 2014 , pages 65-79. Springer, 2014. 5, 6 - source
The Introduction (p.1) states that 'except for certain objects including face, pedestrians and certain rigid objects, detectors do not currently run at video rate' and that the presented architecture detects the 20 PASCAL VOC categories simultaneously at 30Hz, i.e. at video/real-time rate. The Abstract (p.1) and Table 2 (Method) bound the comparison to other DPM speed-up efforts: reference DPM at 0.07Hz, the fastest GPU 'Vanilla DPM' at ~1Hz, FTVQ at 0.53s, and 'over an order of magnitude faster than the previous fastest DPM implementation', so none of the listed variants reaches real-time. Thus our copy supports both parts of the attribution: that this 30Hz DPM runs in real-time and that it is the only such DPM among the speed-up efforts compared. (LLM, conf=0.82)
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 that could be checked couldn’t confirm or dispute this claim.
In the paper
“While it is much faster than R-CNN, it still falls short of real-time and takes a significant accuracy hit from not having good proposals.”
What the sources say
- Could not verifycites [20] [20] K. Lenc and A. Vedaldi. R-cnn minus r. arXiv preprint arXiv:1506.06981 , 2015. 5, 6 - sourceExperimental — accuracy not yet measured
The ConvNet-based Object Detection section identifies R-CNN minus R and says it uses predefined regions in an SPPnet system, but gives no speed comparison with R-CNN. A search for R-CNN minus R and speed found no passage in our copy establishing that it is much faster. (LLM, conf=0.73)
plainstates a cause- 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.
- Could not verify
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“While it is much faster than R-CNN, it still falls short of real-time and takes a significant accuracy hit from not having good proposals.”
What the sources say
- Partly supportedcites [20] [20] K. Lenc and A. Vedaldi. R-cnn minus r. arXiv preprint arXiv:1506.06981 , 2015. 5, 6 - sourceExperimental — accuracy not yet measured
The “ConvNet-based Object Detection” section (page unavailable) says R-CNN minus R uses predefined regions instead of region proposals. It does not say that this causes a significant accuracy loss; the paper’s discussion of low accuracy concerns a different system, naïve Faster R-CNN. (LLM, conf=0.83)
plainstates a cause- 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
84 more claims
Create a free account to see all 121 claims and to verify the 2 references still open.
- Probably not a substantial claim. 17
- It may state two separate claims as one. 15
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 source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“which was not thoroughly investigated in prevIoUs work such as YOLOv1 [20] and YOLOv2 [51] employing MSE/L2 loss (Mean Squared Error).”
From Object Detection in 20 Years: A Survey (2023) · cites this paper as [20]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“the YOLO series models v2-v5 [144-147] are some of the most advanced object detection methods based on DNN, which have good performance in terms of speed and precision.”
From A survey of model compression strategies for object detection (2023) · cites this paper as [144]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“As a typical and more widely used new form of applications [9], various deep learning-based intelligent services and applications have changed many aspects of people's lives due to the great advantages of Deep Learning (DL) in the fields of Computer Vision (CV) and Natural Language Processing (NLP) [10].”
From Convergence of Edge Computing and Deep Learning: A Comprehensive Survey (2020) · cites this paper as [9]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“A convolutional neural network predicts the bounding boxes as well as class probabilities for all the objects depicted in an image.”
From Object detection using YOLO: challenges, architectural successors, datasets and applications (2023) · cites this paper as [56]
The cited source says what the claim attributes to it.
In the citing paper
“Authors of YOLO [56] have reframed the problem of object detection as a regression problem instead of classification problem.”
From Object detection using YOLO: challenges, architectural successors, datasets and applications (2023) · cites this paper as [56]
6 more claims about this paper
Create a free account to see every claim other papers make about this one.
Citations
Cites
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 73,472
Self-citationFaster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2015) · arXiv · doi:10.48550/arxiv.1506.01497 · Cited by 70,812
Self-citationGoing deeper with convolutions
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298594 · Cited by 48,562
ImageNet Large Scale Visual Recognition Challenge
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M., et al. (2015) · International Journal of Computer Vision · doi:10.1007/s11263-015-0816-y · Cited by 39,380 · cited 1×
“We pretrain our convolutional layers on the ImageNet 1000-class competition dataset [29].”
MethodHistograms of Oriented Gradients for Human Detection
Dalal, N., & Triggs, B. (2005) · 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) · doi:10.1109/cvpr.2005.177 · Cited by 33,584 · cited 2×
“Detection pipelines generally start by extracting a set of robust features from input images (Haar [25], SIFT [23], HOG [4], convolutional features [6]).”
MethodRich 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 2×
“Then, classifiers [35, 21, 13, 10] or localizers [1, 31] are used to identify objects in the feature space.”
MethodSelf-citationGirshick, R. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.169 · Cited by 28,105
Self-citationDelving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification
He, K., Zhang, X., Ren, S., & Sun, J. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.123 · Cited by 19,741 · cited 1×
“Convolutional neural networks (CNNs) achieve impressive performance on classification tasks at real-time speeds [13].”
Object recognition from local scale-invariant features
Lowe, D. G. (1999) · Proceedings of the Seventh IEEE International Conference on Computer Vision · doi:10.1109/iccv.1999.790410 · Cited by 16,472 · cited 1×
“Detection pipelines generally start by extracting a set of robust features from input images (Haar [25], SIFT [23], HOG [4], convolutional features [6]).”
MethodRobust 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×
“Detectors for single classes like faces or people can be highly optimized since they have to deal with much less variation [36].”
Background
Cited by
Items 49,901–50,000 of 51,398
Research on a Lossless Downsampling Enhancement Network for UAV-Based Fine Crack Detection
Zhang, T. A., Tao, Y. C., Lu, C., et al. (2026) · doi:10.1109/crsit69981.2026.11605336
“One-stage detectors represented by YOLO have been widely used in real-time object detection because of their efficient end-to-end detection framework [3] .”
A Lightweight Ship Target Detection Algorithm Based on Improved RT-DETR
Zheng, R., & Duan, X. (2026) · International Conference on Electrical, Control and Information Technology (ECITech) · doi:10.1109/ecitech69277.2026.11601881
“Two-stage frameworks, such as Faster R-CNN [10], first generate candidate regions and then perform classification and localization, whereas one-stage detectors, represented by YOLO [11] and SSD [12], directly predict object classes and bounding boxes.”
Vishnu Namboodiri, M., Soni, P., & Joby, N. J. (2026) · doi:10.1109/icati69399.2026.11604290
Super‐Resolution Enhanced Defect Detection Network for Photovoltaic Panel Images
Tang, H., & Yang, B. (2026) · IET Image Processing · doi:10.1049/ipr2.70436
From Recognition to Reasoning: Edge-Frequency Chain-of-Thought for Remote Sensing Object Detection
Li, X., Zhao, Y., Yuan, B., Zhao, Y., & Yao, R. (2026) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2026.3715352
Reddy, N. L., Kapudasi, S., & Mahto, M. K. (2026) · 2026 International Conference on Cognitive Computing and Networking Systems (ICC-CNS) · doi:10.1109/icc-cns70518.2026.11606075
“In order to overcome these limitations, single-stage detectors such as You Only Look Once (YOLO) have been developed, which enable performing the detection faster, by providing a way of processing the whole image in one pass [6].”
Distinguish Traffic Conditions and Suggest Routes Based on YOLO Model and A-star Algorithm
Huu, P. N., Huy, N. D., Quynh, P. T. T., et al. (2025) · Asia Conference on Cognitive Engineering and Intelligent Interaction (CEII) · doi:10.1109/ceii67787.2025.00008
“YOLO [11] offers faster real-time detection and is selected for this study to process videos from Vietnamese intersections.”
Road Damage Detection with an Attention-Enhanced YOLOv8n Model
Yun, F., Duan, W., Hao, Z., et al. (2026) · doi:10.1109/prmvai70103.2026.11605528
“By contrast, one-stage detectors, such as SSD [5] and the YOLO series [6], predict object categories and bounding boxes in a single forward process.”
Research on Interactive Object Detection with Multimodal Large Models
Xiao, H., Chen, X., Xue, H., et al. (2026) · International Conference on Internet of Things, Automation and Artificial Intelligence (IoTAAI) · doi:10.1109/iotaai70471.2026.11607792
“…detectors (e.g., Faster R-CNN) achieve high-precision localization through region proposal networks but suffer from high inference latency; single-stage detectors (e.g., the YOLO series) reformulate detection as a dense regression problem, balancing real-time performance and accuracy [1-2].”
Izadpanah, K., Soheilian, M. H., & Safaei, B. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3715665
Enhanced Garbage Detection Using CNN with EfficientViT and Hybrid Feature Fusion
Kona, M. S. R., Harini, A., Lahari, E. D. S., et al. (2026) · doi:10.1109/icc-cns70518.2026.11605995
Multimodal deep learning for automated atomic qubits fabrication in silicon
Cui, Y., Wu, K., Wu, P., Yang, L., Duan, M., Wang, G., Shi, K., Lyu, Y., Huang, P., Xin, T., et al. (2026) · Machine Learning Science and Technology · doi:10.1088/2632-2153/ae8e30
“The widely used YOLO (You Only Look Once) object detection framework offers a superior alternative to CNN, as it performs real-time localization and classification in a single forward pass [47].”
An improved YOLOv8n vehicle target detection algorithm for adverse weather conditions
Wang, Y., & Song, A. (2026) · International Conference on Algorithms, Software Engineering and Network Security (ASENS) · doi:10.1109/asens69964.2026.11605341
“In contrast, the one-stage algorithm represented by YOLO[2](You Only Look Once) series has achieved a better balance between speed and accuracy due to its end-to-end characteristics.”
YOLOv8-Based Sonar Ship Detection With Noise Feature Enhancement and Support-Guided Adaptation
Tang, Y. (2026) · doi:10.1109/prmvai70103.2026.11605566
“YOLO belongs to the one-stage family and is widely used because of its efficient structure and convenient deployment [1].”
GDW-YOLO: An Insulator Defect Detection Algorithm Based on an Improvement of YOLO11n
Fang, Z., Wu, X., 项顺伯, & Xu, B. (2026) · International Conference on Internet of Things, Automation and Artificial Intelligence (IoTAAI) · doi:10.1109/iotaai70471.2026.11607758
“Current mainstream deep learning algorithms include two-stage detectors like Faster R-CNN [3] and single-stage detectors like SSD [4] and YOLO [5].”
Kothai, G., S, A. K. A., & Reswant Raja, S. (2026) · International Conference on Cognitive Computing and Networking Systems (ICC-CNS) · doi:10.1109/icc-cns70518.2026.11606020
“The computer vision models that have found wide application in vehicle detection in traffic surveillance are YOLO [6] and Faster R-CNN [14].”
The Learning Sentinel: AI-Based Detection of Unauthorized Mobile Usage in Educational Institutions
Bejawada, V., Arepalli, P. S. P., Chithirala, S., et al. (2026) · doi:10.1109/icc-cns70518.2026.11606139
“A single-stage detector named You Only Look Once (YOLO) operates on a single stage. poses object detection as a single regression problem [4].”
Intelligent Detection for Fruit Freshness and Perishability
Yuen, M.-C., & Li, X. (2025) · Asia Conference on Cognitive Engineering and Intelligent Interaction (CEII) · doi:10.1109/ceii67787.2025.00022
“Representative models include YOLO (You Only Look Once) series, SSD (Single Shot MultiBox Detector), and RetinaNet [12].”
InfluentialObject-Level Waste Classification Using EfficientNet-B0 with YOLO-Based Cropping
Parameswara, P. A. Y., Rihonggao, T., Sheza, R. A., Wijayanto, I., & Hadiyoso, S. (2026) · International Conference on Enterprise and Industrial System (ICOEINS) · doi:10.1109/icoeins69996.2026.11605885
“Object-detection algorithms from the YOLO family have demonstrated strong performance in detecting and localizing waste materials in complex and cluttered environments [10].”
Putra, N., Jokonowo, B., Priambodo, B., & Kurniawan, M. I. (2026) · International Conference on Enterprise and Industrial System (ICOEINS) · doi:10.1109/icoeins69996.2026.11605849
Multi-Person Detection and Identity Consistent Crowd Counting in Real Time Video Surveillance
Raj, N., Suresh, P., Babu, S., et al. (2026) · International Conference on Computing, Communication, Security and Intelligent Systems (IC3SIS) · doi:10.1109/ic3sis69949.2026.11608285
“YOLOv8 is a single-stage object detection deep learning-based model that strives to have high detection precision and real-time speed [5], [8].”
InfluentialLiu, R., Wang, R., Wang, Z., et al. (2026) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2026.3715636
Lightweight Detection Method for Welding Crack Defects Based on Improved YOLOv8n
Zhan, Q., & Zhan, T. (2026) · doi:10.1109/prmvai70103.2026.11605861
“In this work, we adopt YOLOv8n [8]as the baseline for lightweight design and apply three targeted “subtractive” operations to reduce model complexity: replacing standard convolutions with lightweight downsampling modules, sparsifying redundant channels through a ghosting mechanism, and integrating…”
SAR Image Object Detection Based on Cascaded Denoising and Counterfactual Self-Distillation
Tian, L., Nie, Y., He, G., Geng, J., Jiang, W., & Jin, S. (2026) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2026.3715503
Vision Model Scaling and Fusion for High-Precision UAV-Based Human Detection
Wang, S. (2026) · International Conference on Internet of Things, Automation and Artificial Intelligence (IoTAAI) · doi:10.1109/iotaai70471.2026.11607805
“Meanwhile, the YOLO series by Redmon et al. [5] pioneered the one-stage detection paradigm, modeling detection as a regression task and striking a good balance between speed and accuracy.”
Baniya, S., Mohanta, T., Yacoub, M., et al. (2026) · Microplastics · doi:10.3390/microplastics5030145
Towards standards-aligned round-timber grading: validation of a semantic segmentation dataset
Bednar, L., Hendrichovsky, F., Motschi, A., Soukup, D., & Meyer-Heye, P. (2026) · European Journal of Wood and Wood Products · doi:10.1007/s00107-026-02443-z
Sheza, R. A., Rihonggao, T., Parameswara, P. A. Y., Wijayanto, I., & Hadiyoso, S. (2026) · International Conference on Enterprise and Industrial System (ICOEINS) · doi:10.1109/icoeins69996.2026.11605897
Image-based Goat Body Size and Weight Estimation via Deep Learning Segmentation and Depth Estimation
Shrestha, S. K., R, H. Y., Theja Sree, M., et al. (2026) · doi:10.1109/icc-cns70518.2026.11606041
“For that field of object detection with the YOLO (You Only Look Once) architecture is proposed, to enable real-time object detection [6].”
Han, S., Yang, M., Li, X., & Wang, K. (2026) · IEEE International Conference on Pattern Recognition, Machine Vision and Artificial Intelligence (PRMVAI) · doi:10.1109/prmvai70103.2026.11605447
“The integration of AI and IoT enables real-time, non-invasive behavioral monitoring, and vision-based fall detection (especially with deep learning models like YOLO) has gained attention for its non-invasiveness and wide coverage [11] .”
DSAF-YOLOv8: A Behavior Detection Algorithm for Dense and Occluded Multi-Person Scenes
Zhang, Y., Chen, Q., Xu, F., Chen, M., Wang, Q. L., & Duan, Z. (2026) · IEEE International Conference on Pattern Recognition, Machine Vision and Artificial Intelligence (PRMVAI) · doi:10.1109/prmvai70103.2026.11605562
“Although the YOLO series has become a mainstream solution because of its favorable speed-accuracy tradeoff and deployment efficiency [4], generic YOLO architectures are primarily optimized for common object detection benchmarks and often struggle in dense behavior-oriented scenes with severe…”
Multi-scale YOLO-MAML for Scaling Few-shot Learning in Agricultural Disease Detection
Liu, B. Z., Zhou, Z., Li, J., & Zhang, Y. (2026) · IEEE International Conference on Pattern Recognition, Machine Vision and Artificial Intelligence (PRMVAI) · doi:10.1109/prmvai70103.2026.11605442
Bing, Y., Xu, J., & Wang, Y. (2026) · doi:10.1109/prmvai70103.2026.11605501
“Among these, the YOLO series has become the mainstream choice for embedded devices due to its fast detection speed, compact architecture and balanced accuracy [1].”
Zhang, X., Tang, H., Hu, B., et al. (2026) · Remote Sensing · doi:10.3390/rs18142422
Enhancing underwater object detection through bidirectional cross-scale gated feature fusion
ANTWI, S. A., Zhou, Z., & Wang, H. (2026) · Engineering Research Express · doi:10.1088/2631-8695/ae8e58
“Since the original YOLO model proposed by Redmon et al [8], successive versions have improved feature extraction, multi-scale representation, and detection efficiency.”
Açıkgöz, Z., & İni̇k, Ö. (2026) · International Symposium on AI-Driven Engineering Systems (ISADES) · doi:10.1109/isades69945.2026.11608072
“Moreover, object detection methods like YOLO have been extensively used to analyze medical images with significant impacts from training strategy and model architecture on performance [12].”
Liu, X. D. (2026) · doi:10.1109/isades69945.2026.11608240
“Standard YOLOv7 + PnP: This refers to the classic deep learning two-dimensional detection combined with the traditional analytic geometry-based pose estimation method [6].”
InfluentialFrom Convolution to Attention and Beyond: A Systematic Review of Modern Vision Architectures
Chakrabarty, A., Aziz, A., Tahsin, A., Nawreen, N. I., Alam, M. G. R., & Maidin, S. S. (2026) · IEEE Access · doi:10.1109/access.2026.3715655
Detection of Wrong Sitting Posture to Prevent Ergonomic Hazards
Singha, A., Sarkar, A., Bhattacharyya, A., Singh, A., Vardhan, M., Parida, M. K., & Sabut, S. (2026) · International Conference on Cognitive Computing and Networking Systems (ICC-CNS) · doi:10.1109/icc-cns70518.2026.11606024 · Cited by 1
“In the real-time keypoint detection pipeline, each video frame captured from the webcam is passed through a YOLOv8 pose estimation model [7] to obtain 2D body keypoints.”
An Automated Bamboo Tube Thickness Measurement System Based on YOLOv11n
Li, Y., Liang, G., Zhou, S., & Kong, Q. (2026) · IEEE International Conference on Pattern Recognition, Machine Vision and Artificial Intelligence (PRMVAI) · doi:10.1109/prmvai70103.2026.11605912
“Furthermore, this single-stage detection strategy perfectly aligns with the fundamental design philosophy of the YOLO series [8].”
Ma, L., Shang, L., Bai, W., et al. (2026) · Engineering Research Express · doi:10.1088/2631-8695/ae8a1c
“Since YOLOv1 [22] redefined the detection task as a regression problem, this series of models has continued to evolve.”
Bhansali, S., Ignatev, N., & Simões, M. (2026) · Natural History Collections and Museomics · doi:10.3897/nhcm.3.196081
Phae, K., Poonsupsopon, P., Dhechasidphaisan, P., et al. (2026) · International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology · doi:10.1109/ecti-con68836.2026.11608974
“…they typically require substantially larger datasets for training and often present higher computational overhead for mobile edge devices compared to optimized CNNs. Similarly, YOLO (You Only Look Once) architectures [12] are currently state-of-the-art for point detection and object bounding.”
BAR-UNet: a boundary-aware and appearance-robust framework for polyp segmentation
Xu, Y., Liu, L., & Tang, M. (2026) · Frontiers in Oncology · doi:10.3389/fonc.2026.1873111
YOLO-Defect: multi-scale feature enhancement for robust bridge surface defect detection
Tang, H., Zhang, X., Liu, Y., Zhai, D., & Li, Y. (2026) · International Journal of Machine Learning and Cybernetics · doi:10.1007/s13042-026-03236-9
“The overall loss function is obtained by summing classification loss, confidence loss, and localization loss, as expressed in equation (9) [ 4].”
InfluentialComparative Study of YOLOv8 and Faster R-CNN in Object Detection
Wu, J. (2026) · IEEE International Conference on Electronic Technology, Communication and Information (ICETCI) · doi:10.1109/icetci68937.2026.11610750
“Each prediction contains location parameters, an object confidence value, and category probabilities [10].”
Parlak, C. (2026) · Tasarım Mimarlık ve Mühendislik Dergisi · doi:10.59732/dae.1827622
Ouyang, J., Du, S., Han, R., Liu, C. H., Zhao, J., Wu, X., & Chen, L. Y. (2026) · International Journal of Computer Vision · doi:10.1007/s11263-026-02969-y
“…models have achieved significant progress in recent years, such as image classification (Krizhevsky et al., 2012; Liu et al., 2025), object detection (Redmon et al., 2016; Chen et al., 2023; Zang et al., 2025), and semantic segmentation (He et al., 2017; Zhang et al., 2023; Yiheng et al., 2023).”
YOLO-OODNet: Attention-Guided Oriented Object Detection for Fish-Eye Imaging Sensors
Zhou, K., Liao, J., Hu, L., Zhang, J., & Peng, S. (2026) · IEEE Sensors Journal · doi:10.1109/jsen.2026.3714195
“In recent years, the YOLO series [33], [34], [35], [36], [37], [38], [39], [40], [41], [42] of models have undergone iterative improvements, achieving significant advancements in both the accuracy and speed of object detection.”
Xia, L., Xu, L., Guo, Y., et al. (2026) · Frontiers in Plant Science · doi:10.3389/fpls.2026.1886272
Vision-Based Multi-Camera Safety Monitoring for Human-Robot Collaboration
Hartanto, J. K., & Budiharto, W. (2026) · IEEE International Conference on Signals and Systems (ICSigSys) · doi:10.1109/icsigsys71160.2026.11613922
“Therefore, an asymmetric tracking architecture is proposed that eliminates point cloud processing, replacing it with YOLO-based object detection, which has been proven efficient for real-time object detection [3, 5].”
Rongxuan, 张. Z., Ke, 孙. S., Hongli, 何. H., et al. (2026) · Chinese Journal of Lasers · doi:10.3788/cjl251705
Benchmark evaluation of unsupervised domain adaptive object detectors in foggy UAV environments
Tran, H., & Nguyen, K.-D. (2026) · International Journal of Machine Learning and Cybernetics · doi:10.1007/s13042-026-03225-y
Lightweight and Maintainable Approach for Table Detection
Venkataraman, V., Shi, L., Gurugnanambiga, S., et al. (2026) · doi:10.1109/icad69378.2026.11608717
“Recent advances in deep learning have substantially improved automated table detection systems, particularly with the emergence of YOLO-based object detectors [1]–[4].”
DA-YOLO: Depth-Aware YOLO framework for Autonomous Driving Using MiDaS
Santhosh, M., Sibi, P., & C, D. J. (2026) · doi:10.1109/icitsif69060.2026.11609063
Jangade, T. (2026) · International Conference on Machine Learning and Intelligent Systems Engineering (MLISE) · doi:10.1109/mlise70044.2026.11607614
“Single-shot detectors including Single Shot MultiBox Detector (SSD) [5] and the YOLO family [1] achieved real-time performance by framing detection as a single regression problem.”
Kailash, N., Janani, S., K, S. J., et al. (2026) · doi:10.1109/icitsif69060.2026.11608841
“The Emergency Vehicle Detection Module uses a combination of visual patterns (e.g., siren lights and colour schemes) and, optionally, audio cues to identify ambulances and fire trucks [11], [12].”
InfluentialImproving Robotic UAV Wind-Turbine Inspection with Digital-Twin Synthetic Data
Silva, J. P. N. Q., Araujo, G. M., Passos, W. L., Barreto, P. S., Santos, T. M. B., & Pinto, M. F. (2026) · Journal of Control Automation and Electrical Systems · doi:10.1007/s40313-026-01312-7
U-Net deep learning model for rapid identification of defects in SiC
Gallagher, J. C., Mahadik, N. A., Stahlbush, R. E., et al. (2026) · Semiconductor Science and Technology · doi:10.1088/1361-6641/ae887e
“The You Only Look Once (YOLO) algorithm [6, 7] processes an entire image within a single neural network to perform object detection, unlike earlier methods that relied on per-region proposals.”
YOLO26: High-Efficiency Object Detection for Plasmodium vivax in Blood Smear Microscopy
Kasamsumran, N., & Pora, W. (2026) · International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON) · doi:10.1109/ecti-con68836.2026.11608776
“YOLOv1(2016)[22] The initial version of YOLO encounters difficulties in identifying small objects and precisely anticipating bounding boxes.”
Radar single-target detection method based on convolutional neural network
Song, Y., Zheng, G., Xiaolong, F., Li, X., & Liu, F. (2025) · doi:10.1109/nnnlp68229.2025.00007
“Models like VGG [4], GoogleNet/Inception [5], YOLO [6], and ResNet [7] all have different design philosophies and network structures, but they have all made significant contributions to the development of convolutional neural networks.”
Akhtamov, A., Ryu, J. H., 박영호, et al. (2026) · Scientific Reports · doi:10.1038/s41598-026-66580-4
People Counting Using YOLO-Based Detection and Clustering for a Mobile Robot
Gomulka, K., Woźniak, P., & Krzeszowski, T. (2026) · Sensors · doi:10.3390/s26154664 · Cited by 2
“However, one of the most well-known models in the literature and used is YOLO [28], renowned for its speed and accuracy, making it particularly suitable for real-time applications.”
CryoAnomaly: Few-Shot Cryo-EM Particle Picking via Anomaly-Guided Hard Negative Suppression
Itsuji, R., Otsubo, R., Fujii, R., Li, X., Wu, X., Saitô, H., & Xu, M. (2026) · IEEE Access · doi:10.1109/access.2026.3716539 · Cited by 1
Mao, N., Ding, S., Zhang, Y. V., et al. (2026) · Sensors · doi:10.3390/s26165185
“Detection frameworks such as YOLO, Faster R-CNN, EfficientDet, and YOLOv4 have promoted research on multiscale feature fusion and hard-example mining strategies for road distress recognition [23–26].”
MWS-DETR: A Lightweight Model for Metal Weldment Surface Small-Scale Defect Detection
Li, Y., & Dm, X. (2026) · IEEE Data Driven Control and Learning Systems Conference (DDCLS) · doi:10.1109/ddcls71227.2026.11610398
“Detection-based frameworks can be further categorized into two-stage detectors, such as Faster R-CNN[10], and one-stage detectors, including the YOLO series[11] and Transformer-based detection models[12].”
Gao, L., Sha, C., Yang, M., Lu, Y., Qiukai, E., & Chen, L. (2026) · Frontiers in Medicine · doi:10.3389/fmed.2026.1875077
TriFusion-DETR: A Light-weight Real-time Detector for Smartphone Screen Glass Defects
Fan, X., Wu, P., Yu, Y., Kandel, N., Yang, Z., & He, G. (2026) · IEEE Data Driven Control and Learning Systems Conference (DDCLS) · doi:10.1109/ddcls71227.2026.11610057
“Currently, YOLO is the mainstream one-stage algorithm [10].”
Allman, M., & Zhu, H. (2026) · Conference proceedings - Canadian Conference on Electrical and Computer Engineering · doi:10.1109/ccece68150.2026.11610223
“Although the YOLO (You Only Look Once) family has advanced state-of-the-art detection performance since its introduction [8], and transformer-based architectures such as DETR have extended detection capabilities to complex visual scenes [9], limited work compares how these models perform on…”
PCCS-YOLOv8: An Enhanced Lightweight Detector for Small UAV Detection in Complex Scenes
Gao, R., Zhang, Y., Xia, Q., & Zhao, Y. (2026) · Electronics · doi:10.3390/electronics15153247 · Cited by 1
SAR Moving Object Detection Method Based on Improved YOLOv8 in Noisy Environment
Li, S., Xinxin, T. A. N. G., & Huang, D. (2026) · doi:10.1109/ddcls71227.2026.11610247
Domain-Adaptive YOLOv9 for Foggy-Weather Object Detection Using Partial Spatial Self-Attention
Xiao, Z., Ma, Y., & Gu, J. (2026) · Conference proceedings - Canadian Conference on Electrical and Computer Engineering · doi:10.1109/ccece68150.2026.11610043
“One-stage detectors in the YOLO family [8] are attractive for practical systems because they offer high throughput with competitive accuracy.”
Habtoul, R., Zaabi, M., & Smaoui, N. (2026) · Network Modeling Analysis in Health Informatics and Bioinformatics · doi:10.1007/s13721-026-00831-2
“This can result in the model becoming biased and performing poorly when applied to new, unseen data (Redmon et al. 2016 ).”
Disentangled Representation Learning for Membership Inference Defense in Graph Neural Networks
Jnaini, A., & Koulali, M. A. (2026) · Conference proceedings - Canadian Conference on Electrical and Computer Engineering · doi:10.1109/ccece68150.2026.11610151
MAD-YOLO: multi-dimensional attention and dynamic feature aggregation for UAV small object detection
Chen, Y., Zhang, S., Deng, Z., Wen, X., Jiang, T., Wu, Y., & Yang, H. (2026) · Measurement Science and Technology · doi:10.1088/1361-6501/ae8f64
“In contrast, one-stage detectors represented by SSD [10] and the YOLO series [11–19] directly predict object categories and bounding boxes in a unified pipeline, providing higher inference efficiency and better real-time performance.”
Lin, H.-Y., Chai, J., Tseng, Q., Lin, C.-W., Pan, K., Liu, K.-W., & Shih, Y.-C. (2026) · Frontiers in Endocrinology · doi:10.3389/fendo.2026.1906679 · Cited by 1
Edge Feature Information Enhancement Algorithm for Small Objects Detection of Remote Sensing Images
Li, H., Tao, H., & Zhuang, Z. (2026) · doi:10.1109/ddcls71227.2026.11610343
“The single-stage al-gorithm includes SSD[10], RetinaNet[11], You Only Look Once (YOLO) [12] and so on.”
Sharma, S. K., & Raheman, H. (2026) · Research Square · doi:10.21203/rs.3.rs-10270873/v1
Misaka, T., Kaminaga, H., Kitamura, T., et al. (2026) · Acta Astronautica · doi:10.1016/j.actaastro.2026.07.054
GDE-YOLO: A Lightweight and High-Precision Small Object Detector for UAV Imagery
李, 淼. (2026) · Computer Science and Application · doi:10.12677/csa.2026.167248
Wang, Y., & Habib, A. (2026) · Cluster Computing · doi:10.1007/s10586-026-06361-w
“…scores ShuffleNet [ 36] 3 3 convolution with batch normalization and activation, followed with a max pooling operation for early downsampling 1 1 pointwise convolution to mix channel information, followed with global average pooling followed by a fully connected layer Tiny YOLO [37] (Detection)”
Structure-Aware Phase-Based Dual Alignment for Robust UAV Object Detection
Baek, M., Oh, H., Yoon, J., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-31441-3_29
Opanasopit, C., & Louis, J. (2026) · Journal of Computing in Civil Engineering · doi:10.1061/jccee5.cpeng-6857
Image Processing Versus Deep Learning for People Counting
Mukai, N., Mirai, A. M. O., & Chang, Y. (2026) · The Journal of the Society for Art and Science · doi:10.3756/artsci.25.7_1
Hashemifard, K., Climent-Pérez, P., & Flórez‐Revuelta, F. (2026) · Journal of Healthcare Informatics Research · doi:10.1007/s41666-026-00244-w
“This has been widely done in egocentric ADL recognition [41]; however, it is not as common in third-person vision-based HAR.”
Guo, L., Gong, J., Zhang, Y., et al. (2026) · IET Cyber-Systems and Robotics · doi:10.1049/csy2.70061
Sweet Corn Maturity Classification using YOLOv7 and ResNet-50
Cadsawan, R. J. R., Figarola, P. J. D., & Caya, M. V. C. (2026) · International Conference on Electronics, Computers and Artificial Intelligence (ECAI) · doi:10.1109/ecai69016.2026.11613735
“As a leading solution for these tasks, the YOLO architecture demonstrates exceptional versatility [11].”
MFIA-YOLO: A Small-Object Defect Detection Model for Transmission Line Spacers
Du, J., Wang, X., Lu, H., et al. (2026) · Energies · doi:10.3390/en19153481
Genetic analyses of leaf traits in an interspecific Zoysia japonica × Zoysia matrella F 2 population
Pradhan, S., Pendergast, T. H., Spiekerman, J. J., Lagergren, J. H., Jacobson, D., & Devos, K. M. (2026) · The Plant Genome · doi:10.1002/tpg2.70282
“Object instances were encoded into a You Only Look Once-style format (Redmon et al., 2016) using a grid size of 16 and five prediction channels (object classification, object center x/y, and object height/width) suitable for training deep learning models.”
The RealDefocus Benchmark for Defocus Deblurring
Seizinger, T., Zhou, Z., & Timofte, R. (2026) · Proceedings - International Conference on Image Processing · doi:10.1109/icip61757.2026.11630040
“While shallow depth of field can be intentionally exploited for artistic purposes [2], unintended defocus or mis-focus frequently degrades image quality and hampers down-stream vision tasks such as text recognition [3, 4], object detection [5, 6], and image classification [7].”
A vision-based AI robotic system for worker safety monitoring in data center environments
Amrani, N. E., En-Najidi, M., kari, B. E., Chaibi, Y., & Chalh, Z. (2026) · International Conference on Electronics, Computers and Artificial Intelligence (ECAI) · doi:10.1109/ecai69016.2026.11613620
“Modern deep learning detectors fall into two broad categories[6]: two stage detectors such as Region-based Convolutional Neural Network (R-CNN) variants[7],which first generate region proposals before classifying them, achieving high accuracy at the cost of real-time performance; and single-stage detectors such as YOL…”
Warehouse Safety and Security Monitoring with Computer Vision
Vimukthi, D., Thilakarathna, P., de Silva, D., et al. (2026) · International Conference on Electronics, Computers and Artificial Intelligence (ECAI) · doi:10.1109/ecai69016.2026.11613798
Chalabi, K., Sabbah, M., Gouget, N. A. D. T., et al. (2026) · The International Journal of Robotics Research · doi:10.1177/02783649261468361
Bouguessa, A., Bekki, I., Daoud, M. A., & Mostefaoui, S. A. M. (2026) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2026.115725
A Hybrid Intelligent Crowd Monitoring System for Real-Time Public Safety
Manikandaraja, G., Jaisanth, M., & Balaji, M. (2026) · Journal of Ubiquitous Computing and Communication Technologies · doi:10.36548/jucct.2026.3.004
“One-stage detectors provided an efficient balance between precision of detection and computational efficiency and were thus well-suited for large scale surveillance applications [1].”
Efficient Action Counting with Dynamic Queries
Ma, X., Li, Z., Shang, Q., Zhu, W., Ci, H., Qiao, Y., & Wang, Y. (2026) · International Journal of Computer Vision · doi:10.1007/s11263-026-02939-4 · Cited by 6
Deep Learning Pose Estimation for Phenotyping of Co‐Occurring Hyperkinetic Movement Disorders
Cif, L., Demailly, D., Horvàth, G. A., et al. (2026) · Annals of Clinical and Translational Neurology · doi:10.1002/acn3.70474
“Feature extraction was implemented in Python (version 3.10.17), leveraging established computational libraries 24–32 to ensure analytical accuracy and methodological reproducibility ( Supplementary Material S4 ).”
InfluentialAI-Based Emission Inventory Modeling: A Comparative Study for Environmental Monitoring
Kayarwar, A. B., Khamborkar, A. K., & Shahare, R. R. (2026) · International Journal of Latest Technology in Engineering Management & Applied Science · doi:10.51583/ijltemas.2026.150600227
Shan, Z., Dou, H., Wu, F., et al. (2026) · Structural Health Monitoring · doi:10.1177/14759217261467676
Enhanced underwater object detection via multi-scale attention and adaptive feature fusion
Chen, H., Yu, Y., & Fu, T. (2026) · Multimedia Systems · doi:10.1007/s00530-026-02563-9
“Although this strategy pro - vides dense supervision signals, accelerates convergence, and improves performance [15], it also introduces multiple overlapping bounding boxes for each target, necessitating manually designed non-maximum suppression (NMS) to remove redundancies.”
Metadata
- Authors
- Redmon, J., Divvala, S., Girshick, R., & Farhadi, A.
- Affiliations
- University of Washington (US)
- Meta (United States) (US)
- Venue
- 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2016-06
- Pages
- 779-788
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W1483870316
- Cited by
- 53,803
- References
- 42
- Citations by year
- 2026: 5347 · 2025: 8912 · 2024: 8953 · 2023: 8264 · 2022: 6785 · 2021: 5741 · 2020: 4223 · 2019: 3286 · 2018: 1560 · 2017: 635 · 2016: 90
- FWCI
- 644.23
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Zenodo (CERN European Organization for Nuclear Research) (cc-by)
- Funders
- National Science Foundation (1338054)
- Office of Naval Research (N00014-13-1-0720, N00014)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Image and Video Retrieval Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Visual Attention and Saliency Detection
- Keywords
- object detection, YOLO, real-time detection, bounding box regression, end-to-end learning, false positive reduction, domain generalization
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.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).
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
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.