What Makes for Effective Detection Proposals?
Open accessThe DOI matches its Crossref record; 69 of the 69 references it cites were found in a registry or an index.
Abstract
Current top performing object detectors employ detection proposals to guide the search for objects, thereby avoiding exhaustive sliding window search across images. Despite the popularity and widespread use of detection proposals, it is unclear which trade-offs are made when using them during object detection. We provide an in-depth analysis of twelve proposal methods along with four baselines regarding proposal repeatability, ground truth annotation recall on PASCAL, ImageNet, and MS COCO, and their impact on DPM, R-CNN, and Fast R-CNN detection performance. Our analysis shows that for object detection improving proposal localisation accuracy is as important as improving recall. We introduce a novel metric, the average recall (AR), which rewards both high recall and good localisation and correlates surprisingly well with detection performance. Our findings show common strengths and weaknesses of existing methods, and provide insights and metrics for selecting and tuning proposal methods.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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
“It is noteworthy that the Recall-to-IoU metric is just loosely [19], [20], [21] related to the ultimate detection accuracy.”
From Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (2017) · 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
“In the case of sliding window based detectors, the imbalance between backgrounds and objects could be as extreme as 10 7 :1 [71].”
From Object Detection in 20 Years: A Survey (2023) · cites this paper as [71]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
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 · cited 1×
“Indeed, there has been some very recent work in this space [60], [61], [62] that shows promising”
BackgroundFaster 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 · cited 1×
“Indeed, there has been some very recent work in this space [60], [61], [62] that shows promising”
BackgroundImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333
Microsoft COCO: Common Objects in Context
Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollár, P., & Zitnick, C. L. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10602-1_48 · Cited by 45,460
Self-citationRich 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
Girshick, R. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.169 · Cited by 28,105
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
Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2015) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2015.2389824 · Cited by 11,971
Object Detection with Discriminatively Trained Part-Based Models
Felzenszwalb, P. F., Girshick, R. B., McAllester, D., & Ramanan, D. (2010) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2009.167 · Cited by 9,853
The Pascal Visual Object Classes Challenge: A Retrospective
Everingham, M., Eslami, S. M. A., Van Gool, L., Williams, C. K. I., Winn, J., & Zisserman, A. (2015) · International Journal of Computer Vision · doi:10.1007/s11263-014-0733-5 · Cited by 7,139 · cited 3×
“We test the generalisation of proposal methods by evaluating on datasets with many additional classes in §4.”
Method
Research works printed but not linked: 7
- [28] A. Humayun, F. Li, and J. M. Rehg, 'Rigor: Recycling inference in graph cuts for generating object regions,' in CVPR , 2014. -
- [22] P. Krähenbühl and V. Koltun, 'Geodesic object proposals,' in ECCV , 2014. -
- [51] P. Dollár, R. Appel, and W. Kienzle, 'Crosstalk cascades for framerate pedestrian detection,' in ECCV , 2012. -
- [59] K. Tang, A. Joulin, L.-J. Li, and L. Fei-Fei, 'Co-localization in realworld images,' in CVPR , 2014. -
- [57] S. Vicente, C. Rother, and V. Kolmogorov, 'Object cosegmentation,' in CVPR , 2011. -
- [54] N. Chavali, H. Agrawal, A. Mahendru, and D. Batra, 'Objectproposal evaluation protocol is 'gameable',' arXiv:1505.05836 , 2015. - · arXiv
- szegedy, 0, Scalable, high-quality object detection, arXiv 1412 1441
Cited by
Items 601–700 of 764
Relief R-CNN : Utilizing Convolutional Feature Interrelationship for Object DetectionPreprint
Li, G., Liu, J., Jiang, C., Zhang, L., Tang, K., & Liu, Y. (2016) · arXiv (Cornell University) · Cited by 2
Identify the cells' nuclei based on the deep learning neural networkPreprint
Zhang, T., & Ma, R. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1911.09830 · Cited by 2
“R-CNN: The Region-based CNN (R-CNN) approach [3] to bounding-box object detection is to attend to a manageable number of candidate object region [2, 4, 5] and evaluated convolutional network independently on each RoI (Regions of Interest).”
MethodAn image-text multimodal fusion of deep learning for detecting insulator defects
Feng, B., Xia, X., Zhang, L., Zhang, W., Xu, Q., Liu, P., & Chen, S. (2025) · International Journal of Parallel Emergent and Distributed Systems · doi:10.1080/17445760.2025.2469512 · Cited by 1
“With the development of detection technology, the composite diagnosis of rotating machinery and a finely tuned method for fault diagnosis using vibrational mode decomposition (VMD) and a convolutional neural network (CNN) was proposed [18].”
Analysis of real-time multi-surveillance detection model using YOLO v5
Pramanik, T., Burade, P., & Sharma, S. (2025) · Indonesian Journal of Electrical Engineering and Computer Science · doi:10.11591/ijeecs.v38.i3.pp1634-1641 · Cited by 2
“A ratio-aware approach which adaptively adjusts the aspect ratio of images was proposed by Hosang et al. [20].”
Raju, A., Indubala, S. V., Krishna, I., Sooraj, S., & Kumar, P. S. (2025) · International Conference on Smart Computing and Communications (ICSCC) · doi:10.1109/icscc66177.2025.11233385 · Cited by 1
“With an average F1-score of 91.6%, the model showed overall good performance and is hence quite efficient for real-time suspicious activity recognition [6].”
InfluentialXu, B., Wu, X., Su, X., & Ke, W. (2026) · Frontiers in Plant Science · doi:10.3389/fpls.2025.1677854 · Cited by 1
“…this study adopts Precision (P), Recall (R), mean Average Precision (mAP), and mAP50 – 95 as the main evaluation metrics (Everingham et al., 2010; Hosang et al., 2016), de fi ned as follows: (1) Precision (P): re fl ects the proportion of samples predicted by the model to belong to positive…”
Feature enhancement and fusion-optimized defect detection model for Sanhua plums
Chen, Y., & Zhang, Y. (2026) · Frontiers in Plant Science · doi:10.3389/fpls.2026.1759809 · Cited by 1
“This study employed four core metrics to construct an evaluation framework: Precision (P), Recall (R), Mean Average Precision (mAP), and Average Precision across thresholds (mAP50 – 95) (Everingham et al., 2010; Hosang et al., 2016).”
Sichkar, V. N., & Mouromtsev, D. I. (2026) · International Conference on Control and Computer Vision (ICCCV) · doi:10.1145/3810417.3810419 · Cited by 1
Object proposals using SVM-based integrated model
Geng, W., Li, S., Ren, T., & Wu, G. (2016) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2016.7727741 · Cited by 1
“More details about the proposal methods can be found in the recent survey papers [18].”
BackgroundBandwidth Limited Object Recognition in High Resolution Imagery
Lopez-Fuentes, L., Bagdanov, A. D., van de Weijer, J., & Skinnemoen, H. (2017) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2017.138 · Cited by 1
“These proposals may be generated by Selective Search [14], Edge Boxes [17], or any bounding box proposal strategy one prefers [6].”
BackgroundLi, G., Liu, J., Jiang, C., & Tang, K. (2016) · Cited by 2
Semantic context and depth-aware object proposal generation
Zhang, H., He, X., Porikli, F., & Kneip, L. (2016) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2016.7532307 · Cited by 1
“Generating object proposals has become a critical step in topperforming object detection systems [1, 2, 3], which helps reduce the search space of detection to a relatively small number of interesting regions [4].”
BackgroundA chest-shape target automatic detection method based on Deformable Part Models
Zhang, M., Jin, W., & Li, L. (2016) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.2245832 · Cited by 1
An advanced local offset matching strategy for object proposal matching
Zhang, R., & Wang, W. (2016) · Visual Communications and Image Processing · doi:10.1109/vcip.2016.7805495 · Cited by 1
“Four different matching algorithms (NAM, PHM, LOM, ALOM) are compared with various object proposals and features.”
MethodLearning to Generate Object Segment Proposals with Multi-modal Cues
Zhang, H., He, X., & Porikli, F. (2017) · Lecture notes in computer science · doi:10.1007/978-3-319-54181-5_8 · Cited by 1
Lazy dragging: effortless bounding-box drawing for touch-screen devices
Kim, K., Kim, S., & Sohn, K. (2017) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2017.014769 · Cited by 1
“In the study of detection proposals [14], [15], Structural Edges (SE) is frequently utilized [5] to generate detection proposals that are very likely to contain objects.”
BackgroundObject Proposal Generation Applying the Distance Dependent Chinese Restaurant Process
Lauri, M., & Frintrop, S. (2017) · Lecture notes in computer science · doi:10.1007/978-3-319-59126-1_22 · Cited by 1
Adaptive video object proposals by a context-aware model
Geng, W., Zhang, C., & Wu, G. (2017) · Multimedia Tools and Applications · doi:10.1007/s11042-017-4561-9 · Cited by 1
“For most video object proposals start from image object proposals [27, 40, 56], there are generally two categories of object proposals for both image and video, segment-based proposals [3, 51] and window-based proposals [24, 32].”
BackgroundMethodObject Detection with Proposals in High-Resolution Optical Remote Sensing Images
Ding, H., Luo, Q., Zou, Z., Guo, C., & Shi, Z. (2017) · Lecture notes in computer science · doi:10.1007/978-3-319-68935-7_27 · Cited by 1
Accurate object matching for UAV imagery using multi-scale best-buddies similarity
Xu, L., Wang, J., Fu, K., Yu, H., & Yang, W. T. (2017) · IEEE International Geoscience and Remote Sensing Symposium (IGARSS) · doi:10.1109/igarss.2017.8127713 · Cited by 1
Insights of object proposal evaluation
Wang, Y., Huang, L., Ren, T., Zhong, S.-H., Gu, H., & Liu, Y. (2017) · Multimedia Tools and Applications · doi:10.1007/s11042-017-5471-6 · Cited by 1
Saliency Aggregation: Multifeature and Neighbor Based Salient Region Detection for Social Images
Liang, Y., Lang, C., Yu, J. F., Liu, H., & Ma, N. (2018) · Applied Computational Intelligence and Soft Computing · doi:10.1155/2018/1014595 · Cited by 1
“In addition, various priors are also very important in salient region detection [13], for example, face [14–16], car [17], color [14], center bias [13], and objectness [18–20].”
BackgroundZhou, S., & Wei-hua, L. (2017) · International Conference on Computer Systems, Electronics and Control (ICCSEC) · doi:10.1109/iccsec.2017.8446790 · Cited by 1
“The essence of deep learning is used to learn more useful features by building a machine learning model with a lot of hidden layers and massive training data to ultimately improve the accuracy of classification or prediction [12, 13].”
MethodObject Proposal Generation for Unsupervised Object Localization
Guo, W., Lin, R., Wang, S., & Xiong, N. N. (2018) · International Symposium on Parallel Architectures, Algorithms and Programming · doi:10.1109/paap.2018.00046 · Cited by 1
“The key aspect of this proposal method is to score these windows through multiple cues including Multiscale Saliency (MS), Color Contrast (CC), Edge Density (ED), Superpixels Straddling (SS) and Location and Size (LS) [2], [5], [13].”
BackgroundMethodInfluentialAn improved YOLOv2 model with depth-wise separable convolutional layers for object detection
Han, Z., Wang, D., Men, A., & Zhou, Y. (2019) · International Conference on Graphics and Image Processing (ICGIP) · doi:10.1117/12.2524181 · Cited by 1
Scene-Guided Region Proposal Re-ranking Method for On-road Vehicle Candidate Generation
Nan, Z., Feng, Y., He, J., Wei, P., Xu, L., Sun, H., & Zheng, N. (2019) · IEEE Intelligent Vehicles Symposium · doi:10.1109/ivs.2019.8814149 · Cited by 1
“As a result, the sliding window strategy typically generates 10(4)-10(5) candidates on a single-scale image, the number increases to 10(6)-10(7) when considering multiple-scale images [15].”
BackgroundDeep Learning-Based Scene Image Detection and Segmentation with Speech Synthesis in Real Time
Stephen, O., & Sain, M. (2020) · Intelligent systems reference library · doi:10.1007/978-3-030-37551-5_10 · Cited by 1
Application of Neural Network in Image Detection of Illegal Billboards
Jiang, X.-H., Feng, H.-L., & Dong, Y. (2020) · International Academic Conference on Frontiers in Social Sciences and Management Innovation (IAFSM) · doi:10.2991/assehr.k.200207.003 · Cited by 1
“In [2], 12 common methods for generating candidate regions were evaluated in depth.”
MethodModel transferring for unsupervised object localization
Lin, R., Wang, S., & Guo, W. (2019) · IEEE Conference on Cybernetics and Intelligent Systems · doi:10.1109/cis-ram47153.2019.9095838 · Cited by 1
“It is worth noting that object proposal generations could be regarded as two types: grouping methods and window scoring methods [17].”
BackgroundImproved YOLO model with multi-feature fully convolutional network for object detection
Chen, Y., Wang, H., & Han, Z. (2020) · International Workshop on Pattern Recognition (IWPR) · doi:10.1117/12.2574417 · Cited by 1
“Cascade layer [11] is used to fuse multi-scale feature map into a whole.”
MethodDas, K., & Baruah, A. K. (2021) · Innovations in Energy Management and Renewable Resources (IEMRE) · doi:10.1109/iemre52042.2021.9386732 · Cited by 1
Live Video Action Recognition from Unsupervised Action Proposals
Lopcz-Sastrc, R. J., Baptista-Ríos, M., Acevedo-Rodríguez, F. J., Martin-Martin, P., & Maldonado-Bascón, S. (2021) · doi:10.23919/mva51890.2021.9511355 · Cited by 1
“Driven by the success of the object proposals concept for detecting objects in still images [15], some authors proposed to extend this idea to the TAL problem, e.”
BackgroundContent modification of soccer videos using a supervised deep learning framework
Ghassab, V. K., Maanicshah, K., Green, P. E., & Bouguila, N. (2021) · Multimedia Tools and Applications · doi:10.1007/s11042-021-11383-0 · Cited by 2
Miao, R., Jiang, H., Tian, F., & Li, X. (2022) · The Journal of Supercomputing · doi:10.1007/s11227-022-04400-y · Cited by 1
“Some researchers [9–13] proposed a salient feature description to detect small targets to determine small infrared targets [14].”
BackgroundUFO RPN: A Region Proposal Network for Ultra Fast Object Detection
Li, W., & Song, A. (2022) · Lecture notes in computer science · doi:10.1007/978-3-030-97546-3_50 · Cited by 1
Study of Security and Privacy Challenges and their solutions in IoT data
(2021) · American Journal Of Advanced Computing · doi:10.15864/ajac.2102 · Cited by 1
Study of Deep Convolutional Neural Network for Vehicle Localization on Blurred Aerial Imagery
Ilina, O. V., & Tereshonok, M. V. (2022) · 2018Systems of Signal Synchronization, Generating and Processing in Telecommunications (SYNCHROINFO) · doi:10.1109/synchroinfo55067.2022.9840965 · Cited by 1
Building detection based on a boundary-regulated network and watershed segmentation
Zhuang, L. U. O., Li, M., & ZHANG, D. (2022) · National Remote Sensing Bulletin · doi:10.11834/jrs.20219335 · Cited by 2
A Hierarchical Cluster Tree Approach Leveraging Delaunay Triangulation
Avatavului, C., & Boiangiu, C.-A. (2023) · BRAIN BROAD RESEARCH IN ARTIFICIAL INTELLIGENCE AND NEUROSCIENCE · doi:10.18662/brain/14.3/482 · Cited by 1
“Linear NMS, introduced by Hosang et al. (2016), offers a more flexible approach by gradually decreasing the scores of suppressed boxes based on their overlap with the retained box.”
BackgroundBadminton Detection Using Lightweight Neural Networks for Service Fault Judgement
Li, T., Lin, J., Pan, L., & Wang, Z. (2024) · Communications in computer and information science · doi:10.1007/978-981-97-2275-4_15 · Cited by 1
Robust Single-Cam Surround View Object Detection and Localization Using Memory Maps
Quan, Y., Kiefer, B., Messmer, M., Akupati, C. R., Graser, R., & Zell, A. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-78113-1_8 · Cited by 1
Fuzzy Inference based Pedestrian and Cyclist Tracking in a Compressed Video
Praveenkumar, S. M., Patil, P. R., Karadgi, S., & Hiremath, P. S. (2024) · IEEE Conference on Engineering Informatics (ICEI) · doi:10.1109/icei64305.2024.10912237 · Cited by 1
“In the literature, there are many methods for object recognition, e.g., segmentation clustering-based method and window scoring based method [16].”
Jin, X., Fu, H., Kong, J., Ma, H., Bai, Y., & Su, T. (2025) · Chinese journal of information fusion. · doi:10.62762/cjif.2025.919344 · Cited by 1
Diversity in Object ProposalsPreprint
Winschel, A., Lienhart, R. W., & Eggert, C. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1603.04308 · Cited by 1
“Individually, SS and EB yield comparable recall results in many cases (see [7] for details).”
BackgroundRobust Anomaly-Based Ship Proposals Detection from Pan-sharpened High-Resolution Satellite ImagePreprint
Luu, V. H., Luong, N. H. H., Bui, Q. H., & Nguyen, T. N. T. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1804.09322 · Cited by 1
Effective and annotation efficient deep learning for image understanding
Gidaris, S. (2018) · theses.fr (ABES) · doi:10.70675/536a95d2zabb4z4063z9aa6z8d6dd534ca1e · Cited by 1
“[64] that includes a sliding window implementation inspired by BING [14, 173]).”
BackgroundMethodInfluentialOnline deep transfer learning applied to building quality assessment robots
Liu, L. (2019) · DR-NTU (Nanyang Technological University) · doi:10.32657/10356/104468 · Cited by 2
A Study on Evaluation Standard for Automatic Crack Detection Regard the Random FractalPreprint
Li, H., Wang, J., Zhang, Y., Wang, Z., & Wang, T. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2007.12082 · Cited by 2
“However, there are controversies about this standard in general object detection [23, 24], that the mAP scores could not indicate the visual performance of the detection reasonably.”
BackgroundObject Recognition Using Deep Learning in IoT ApplicationsPreprint
Yadav, P., & Kumari, M. (2018) · SSRN Electronic Journal · doi:10.2139/ssrn.3168605 · Cited by 1
Tang, Y. (2016) · theses.fr (ABES) · doi:10.70675/a976a56czff7dz452bz99c2zc569c9957707 · Cited by 1
Intrinsic motivation mecanisms for incremental learning of visual saliency
Craye, C. (2017) · theses.fr (ABES) · doi:10.70675/922227cfzcecaz436azb9fez23e8903dbf49 · Cited by 3
“[172] proposed to use the average Recall (AR) that uses an integral over a set of threshold and demonstrated that this approach was more objective and stable over datasets.”
BackgroundMethodBenchmark evaluation of object segmentation proposal
Hasan, I. (2015)
Investigations of an "Objectness" Measure for Object Localization
Coates, L. (2000) · doi:10.15760/etd.2944
“Further reading on this can be seen in[7]”
BackgroundDouble Layer Multi-Threshold Extension improving the generic object proposal generation
Chen, X., Zhang, C., & Yan, F. (2016) · Chinese Control Conference (CCC) · doi:10.1109/chicc.2016.7554472
“For adequately demonstrating our proposed approach, following [25] and [26], the performance is mainly measured in line with the detection recall (DR) and average recall (AR).”
BackgroundRelief R-CNN: Utilizing Convolutional Feature Interrelationship for Fast Object Detection DeploymentPreprint
Li, G., Liu, J., Jiang, C., Zhang, L., Tang, K., & Zhu, Z. (2016) · arXiv (Cornell University)
Proposal Generating Method Using Geometrical Features for Vehicle Detection
Cai, W., Liang, Z., Liu, X., & Liu, H. (2016) · Lecture notes in electrical engineering · doi:10.1007/978-981-10-2338-5_53
Improving object detection via improving accuracy of object localization
Zhao, H., Jiang, Z., & Men, A. (2016) · IEEE/CIC International Conference on Communications in China (ICCC) · doi:10.1109/iccchina.2016.7636786
“Specifically, [27] compares the influence of different object proposals to R-CNN work like Table 1 shows.”
BackgroundMethodInfluentialA trajectory-based approach for object detection from video
Jiang, C., Li, G., Liu, J., Liu, Y., & Tang, K. (2016) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2016.7727564
“…in original Fast R-CNN as proposal generator, Edge Boxes [20] is also an excellent proposal extract method; According to the experimental results of [21], Edge Boxes can achieve almost the same performance as Selective Search while can be much faster than Selective Search by more than 30 times.”
MethodA comparative study of object proposals re-ranking methods for object detection
Chen, S., Li, J., Hu, X., & Zhou, P. (2016) · International Conference on Wireless Communications and Signal Processing (WCSP) · doi:10.1109/wcsp.2016.7752644
“[12] have provided a deeply analysis of existing proposal generation methods and their impact on the performance of RCNN, and Fast R-CNN, there are still lacking an analysis of proposal re-ranking.”
BackgroundMethodRecognition and detection of objects using visual and textual cues
Karaoğlu, S. (2016) · UvA-DARE (University of Amsterdam)
Object Proposals in Computer Vision
Chavali, N. (2015) · VTechWorks (Virginia Tech)
Object proposal method based on integration of geometrical features
Cai, W., Feng, Y., Liu, X., & Liu, Y. (2016) · International Congress on Image and Signal Processing (CISP) · doi:10.1109/cisp-bmei.2016.7852767
“Currently, object proposal approaches can be separated into two classes: grouping approaches and scoring approaches [20].”
BackgroundUnderstanding Visual Appearance On The Web Using Large-Scale Crowdsourcing And Deep Learning
Bell, S. (2016) · eCommons (Cornell University) · doi:10.7298/x4348h9v
“Following previous approaches [66], we use object proposal detectors [80, 177, 205] to identify ROIs in an image.”
MethodA two-stage cascaded foreground seeds generation for parametric min-cuts
Li, S.-M., Zhu, J.-G., Gao, C., & Li, C.-W. (2016) · KSII Transactions on Internet and Information Systems · doi:10.3837/tiis.2016.11.020
“O bject Proposal [1-4] is a fast object location method.”
MethodRole of Premises in Visual Question Answering
Mahendru, A. (2017) · VTechWorks (Virginia Tech)
Structural sparse representation for object detection
Fang, W., Chen, J., & Hu, R. (2017) · Wuhan University Journal of Natural Sciences · doi:10.1007/s11859-017-1253-2
Environment perception in racing simulators using Deep Neural Networks
Marina, L., Moldoveanu, F. D., & Grigorescu, S. M. (2017) · International Aegean Conference on Electrical Machines and Power Electronics & International Conference on Optimization of Electrical and Electronic Equipment (ACEMP-OPTIM) · doi:10.1109/optim.2017.7975119
“There is a large literature that covers this subject [5], [6], [7].”
BackgroundSaliency detection based object proposal
Azaza, A., & Douik, A. (2017) · International Multi-Conference on Systems, Signals & Devices (SSD) · doi:10.1109/ssd.2017.8167000
“These methods introduce a set of segments which contain an object [5].”
MethodA Generation Method of Insulator Region Proposals Based on Edge Boxes
Zhao, Z., Zhang, L., Qi, Y., & Shi, Y. (2017) · Communications in computer and information science · doi:10.1007/978-981-10-7299-4_19
Pedestrian proposal generation using depth-aware scale estimation
Park, K., Kim, S., & Sohn, K. (2017) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2017.8296641
“Comparison with the state-of-the-art algorithms Following qualitative experiments reported in [23, 6, 10], we measured the performance of proposal algorithms such as SS [12], EB [9], MCG [11], and 3DOP [6] including ours as shown in Fig.”
MethodFusion of Modern and Tradition: A Multi-stage-Based Deep Network Approach for Head Detection
Hsu, F., & Hung, C. (2018) · Lecture notes in computer science · doi:10.1007/978-3-319-93034-3_32
An extension of BING to high IOU threshold
Guo, C., & Zhan, Y. (2018) · International Workshop on Pattern Recognition (IWPR) · doi:10.1117/12.2501932
“Besides, BING is strongly robust to the change of illumination, rotation and scale than most of other algorithms according to the reproducibility of standard provided by Jan Hosang[18].”
MethodEpistemic guidance of visual attention for robotic agents in dynamic visual scenes
Bhakta, A. (2018) · Open Access Victoria University of Wellington | Te Herenga Waka (Figshare) · doi:10.26686/wgtn.17072006
Feature Synthesization for Real-Time Pedestrian Detection in Urban Environment
Fang, W., Chen, J., Lü, T., & Hu, R. (2018) · Lecture notes in computer science · doi:10.1007/978-3-030-00767-6_10
Weight parameterizations in deep neural networks
Zagoruyko, S. (2018) · theses.fr (ABES) · doi:10.70675/90fb1140zc5eaz4aadzbcb7zfd89ca9e7adf
Low altitude polarization hyperspectral target detection based on CNN method
Cao, Y., Xu, G.-M., & Xu, M. (2018) · Conference on Frontiers in Optical Imaging Technology and Applications · doi:10.1117/12.2506662
“In view of the problems existing in the early framework, Hosang et al [7] put forward a solution called region proposal (RP).”
BackgroundDetection proposal method based on shallow feature constraints
Chen, H., Zheng, H., & Deng, Y. (2018) · International Conference on Image Processing Theory, Tools and Applications (IPTA) · doi:10.1109/ipta.2018.8608148
“of-the-art detection proposal methods [4][5], to justify the performance of this method.”
BackgroundMethodInfluentialUnsupervised Object Proposal Using Depth Boundary Density and Density Uniformity
Hosono, T., Tarashima, S., Shimamura, J., & Kinebuchi, T. (2018) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2018.8594408
“Since a good object proposal algorithm should generate accurate candidate regions with a short running time and with a small number of candidates [3], we evaluated these methods in terms of running time and recall versus number of proposals.”
MethodInfluentialLocal Computation with Adaptive Spatial Clustering for Multi-Size Motion Patch Proposals in WAMI
Lin, Y., Xia, Y., Wu, F., & Zhao, J. (2018) · IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS) · doi:10.1109/avss.2018.8639111
“Traditional object proposals utilize motion-based segmentation algorithms [15] to output irregular blobs as objects, or locate targets by single image feature (color, texture or shape)[5].”
BackgroundOrthogonal object proposal and its application
Park, S., & Kwon, J. (2019) · IET Computer Vision · doi:10.1049/iet-cvi.2018.5346
“Hosang et al. [16] extensively surveyed traditional object proposal methods not based on deep learning.”
BackgroundA joint mechanism algorithm for object detection
Han, Z., Wang, D., Men, A., & Zhou, Y. (2019) · International Conference on Graphics and Image Processing (ICGIP) · doi:10.1117/12.2524207
Segmentation Free Object Discovery in VideoPreprint
Cuffaro, G., Becattini, F., Baecchi, C., Seidenari, L., & Bimbo, A. D. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1609.00221
“Object proposals [3] provide a relatively small set of bounding boxes likely to contain salient regions in images, based on some objectness measure.”
BackgroundAddressing ambiguity in multi-target tracking by hierarchical strategy
Taalimi, A., Liu, L., & Qi, H. (2017) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2017.8296960
“…three systems are reported: CR-RMOT: RMOT[27] with ConfRank; CR-ODAL: TC ODAL[15] with ConfRank; CRnop: TVOD with visual tracker without detector [22]; Comparing CR-RMOT and CR-ODAL with RMOT and TC ODAL shows that when measured by MOTA, the most comprehensive metric for tracking performance,…”
BackgroundMethodA Fast Object Proposal Assessment Method Based on Edge Directions
Tan, K., Li, W., Wu, Q., Meng, F., & Xu, L. (2018) · IEEE International Conference on Computer and Communications (ICCC) · doi:10.1109/compcomm.2018.8780841
“In this paper, we report the most commonly used the recall to evaluate the quality of object proposals [28].”
MethodInfluentialA Machine Learning Based Image Inpainting System on Mobile Device
Wang, X., Feng, J., Tang, R., Wang, Z., Tian, C., Tian, Y., & Gong, X. (2019) · International Conference on Artificial Intelligence and Computer Science (AICS) · doi:10.1145/3349341.3349484
Perceptually motivated algorithm for grouping objects in complex scenes
Cheng, Y., Wang, Z., Xu, G., Han, X., & Wang, Z. (2019) · Journal of Electronic Imaging · doi:10.1117/1.jei.28.5.053006
A Lightweight Region Proposal Network for Task Specific Applications
O’Keeffe, S., & Villing, R. C. (2019) · Irish Signals and Systems Conference · doi:10.1109/issc.2019.8904949
“In standard region proposal networks, regions are proposed with bounding box coordinates and a score indicating how likely it is that the bounding box contains an object [4].”
BackgroundJoint boundary-aware method and Bayesian Inference for object proposal
Zhao, H., Chen, Z., Wang, J., Dai, D., & Bao, P. (2019) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2019.11.047
МОДЕЛЬ КАСКАДУ ГІБРИДНОЇ СЕГМЕНТАЦІЇ ДЛЯ АВТОМАТИЧНОГО ДЕШИФРУВАННЯ ОБ’ЄКТІВ НА АЕРОЗНІМКАХ
Kovbasiuk, S., Kanevskyy, L., & Romanchuk, M. P. (2019) · Сучасні інформаційні технології у сфері безпеки та оборони · doi:10.33099/2311-7249/2019-35-2-65-70
Application of Recurrent Convolution Neural Network for Vehicle Detection
Chate, M., Mohanty, M. N., & Gohokar, V. V. (2019) · International Conference on Applied Machine Learning (ICAML) · doi:10.1109/icaml48257.2019.00031
“In [1], [10], [11] can be found complete investigations and comparison of proposed object methods.”
BackgroundApplying SSD to Real World Food Production Environments
Klaeger, T., Schroth, M., Schult, A., & Oehm, L. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-45096-0_33
A High-Performance Object Proposals based on Horizontal High Frequency SignalPreprint
Jiang, C., Huawei, L., & Zhiling, W. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2003.06124
“Pds (m,n) = [2 , 2], { if m ∈ [0, 1, 2, 3], |n−m| ≤ 2”
BackgroundMethodInfluentialSupervision Levels Scale (SLS)Preprint
Damen, D., & Wray, M. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.09890
“Previous efforts to find discrete scales or categories have focused on identifying easy or hard examples within one datasets, in order to provide insights into methods failure cases (e.g. [1,9,10]).”
Classification d'événements à partir de capteurs sols - Application au suivi de personnes fragiles.
Minvielle, L. (2020) · theses.fr (ABES) · doi:10.70675/9ec8e913ze129z428czbbf1z3dc9a2d3310f · Cited by 2
Visual Tracking of Small Unmanned Aerial Vehicles Based on Object Proposal Voting
Hong, J., & Kwon, J. (2021) · Applied Sciences · doi:10.3390/app11030953
“[14] summarized the advantages of several object proposal methods.”
MethodComparison of Region Proposal Methods for Marine Holograms
Liu, Z., Takahashi, T., Thevar, T. M., LINDSAY, D. J., Burns, N., Watson, J., & Thornton, B. (2020) · Global Oceans 2020: Singapore – U.S. Gulf Coast · doi:10.1109/ieeeconf38699.2020.9389216
Common visual pattern discovery and analysis
Wang, Z. (2021) · DR-NTU (Nanyang Technological University) · doi:10.32657/10356/146898
Selective Deep Convolutional Framework for Vehicle Detection in Aerial Imagery
Sakhare, K. V., & Vyas, V. (2021) · Transactions on computer systems and networks · doi:10.1007/978-981-16-1681-5_23
Li, S., Wang, T., Yang, S., & Yuan, Y. (2021) · International Conference of Optical Imaging and Measurement (ICOIM) · doi:10.1109/icoim52180.2021.9524402
“The structure is shown in Fig.2 We follow encoder-decoder architecture like Efficient ConvNet and ENet [15, 16, 22].”
MethodHow Well Can a CNN Marginalize Simple Nuisances It is Designed forPreprint
Karianakis, N., Dong, J., & Soatto, S. (2015) · arXiv (Cornell University)
Metadata
- Authors
- Hosang, J., Benenson, R., Dollar, P., & Schiele, B.
- Affiliations
- Max Planck Institute for Informatics (DE)
- ORCID
- Schiele, Bernt
- Venue
- IEEE Transactions on Pattern Analysis and Machine Intelligence
- Published
- 2016-04-01
- Volume
- 38
- Issue
- 4
- Pages
- 814-830
- ISSN
- 0162-8828 (ISSN-L)
- 2160-9292
- OpenAlex
- W1958328135
- Cited by
- 727
- References
- 68
- Citations by year
- 2026: 9 · 2025: 27 · 2024: 34 · 2023: 36 · 2022: 57 · 2021: 58 · 2020: 63 · 2019: 95 · 2018: 93 · 2017: 126 · 2016: 98
- FWCI
- 100.45
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Transactions on Pattern Analysis and Machine Intelligence (publishedVersion)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- MPG.PuRe (Max Planck Society) (submittedVersion, other-oa)
- MPG.PuRe (Max Planck Society) (submittedVersion)
- arXiv (Cornell University)
- 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, PASCAL dataset, ImageNet, MS COCO, DPM, R-CNN, Fast R-CNN, sliding window search
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.
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.