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 701–764 of 764
Layout Analysis for Handwritten Documents. A Probabilistic Machine Learning Approach
Díaz, L. Q. (2022) · RiuNet (Universitat Politècnica de València) · doi:10.4995/thesis/10251/181483
An empirical study on fall-detection using K-means based training data subset selection
(2021) · American Journal Of Advanced Computing · doi:10.15864/ajac.2101
Mobile Application Development for West Bengal Tourism
(2021) · American Journal Of Advanced Computing · doi:10.15864/ajac.2105
Prime Generation: Algorithms and Analyses
(2021) · American Journal Of Advanced Computing · doi:10.15864/ajac.2108
A Review on SQL Injection Attack in Internet of Things
(2021) · American Journal Of Advanced Computing · doi:10.15864/ajac.2106
Computer Vision Based Pre-Processing System for Autonomous Vehicles
(2021) · American Journal Of Advanced Computing · doi:10.15864/ajac.2103
Locating Glass Insulator Region Proposals of Aerial Images Based on HSV-EdgeBoxes
栾, 乐. (2019) · Computer Science and Application · doi:10.12677/csa.2019.911233
(2021) · American Journal Of Advanced Computing · doi:10.15864/ajac.2107
Computer Vision-based Monitoring of Harvest Quality
Rasmussen, C. B. (2021) · VBN Forskningsportal (Aalborg Universitet) · doi:10.54337/aau468596652
ENHANCED DEEP TEMPLATE-BASED OBJECT INSTANCE DETECTION
Frink, T. (2022) · Digital Commons - URI (University of Rhode Island) · doi:10.23860/thesis-frink-travis-2022
Parameter Investigation in Low Computing Cost Model-Based EfficientDet for UAV Object Detection
Pramawijaya, I. N., Wibowo, S. A., & Usman, K. (2022) · International Conference on Information and Communications Technology (ICOIACT) · doi:10.1109/icoiact55506.2022.9971858
“In many scenario, decreases of the detection quality to prioritise detection speed is common and tolerable [30].”
BackgroundMethodLukyanov, A. D., Aznabaev, M. I., Plekhanova, E. V., Kapitanova, I. R., & Plekhanova, T. G. (2022) · H&ES Research · doi:10.36724/2409-5419-2022-14-5-45-52
MTUCI, Шагров, Н. А., Bolshakov, A. V., & MTUCI (2022) · H&ES Research · doi:10.36724/2409-5419-2022-14-5-29-37
FREQUENCY STABILIZATION BASED ON PRIMARY FUNDAMENTAL PROPERTIES OF LARGE SYSTEMS
Safaryan, O. A., Alferova, I., Naydenova, J. I., & Reshetnikova, I. V. (2022) · H&ES Research · doi:10.36724/2409-5419-2022-14-5-53-58
ON THE IDENTIFICATION OF THE ATTENUATION MODEL IN THE PROPAGATION OF RADIO WAVES
Golovskoy, V. A. (2022) · H&ES Research · doi:10.36724/2409-5419-2022-14-5-38-44 · Cited by 2
Shvidchenko, S. A., Ivanov, S. V., Khorolsky, E. M., & Saveliev, I. V. (2022) · H&ES Research · doi:10.36724/2409-5419-2022-14-5-21-28 · Cited by 2
A DEEP CONVOLUTIONAL NEURAL NETWORK TRAINING TECHNIQUE FOR REMOTE SENSING ROBUST OBJECT LOCALIZATION
Ilina, O. V., & Tereshonok, M. V. (2022) · H&ES Research · doi:10.36724/2409-5419-2022-14-5-4-13
Hou, Y., Wang, Y., Fu, Z., & Zhou, S. (2022) · International Conference on Robotics and Automation Engineering (ICRAE) · doi:10.1109/icrae56463.2022.10056180
“Some work starts by selecting areas with high repetitiveness [18].”
BackgroundMethodA REVIEW OF CROWN DETECTION IN YAKSHAGANA IMAGES USING CNN ALGORITHM
(2023) · International Research Journal of Modernization in Engineering Technology and Science · doi:10.56726/irjmets40321 · Cited by 1
FACE RECOGNITION TECHNOLOGY IMPROVING USING CONVOLUTIONAL NEURAL NETWORKS
Anufriiev, P. (2022) · Scientific papers of Donetsk National Technical University Series Informatics Cybernetics and Computer Science · doi:10.31474/1996-1588-2021-2-33-70-77
PGN: A High Quality Proposal Generation Network for Object Detection
Wang, Y., Qu, F., Wang, X., Xu, Y., & Li, Z. (2023) · International Conference on Frontiers of Electronics, Information and Computation Technologies · doi:10.1109/icfeict59519.2023.00100
“According to Hosang's research [11], the average recall(AR) in the object proposal approach correlates with the precision of final detection linearly, in which AR essentially shows how accurate the predicted bounding boxes are localized compared with the ground truth.”
Research on Lightweight Network Segmentation of Lung Parenchyma CT Image Based on FPGA Platform
Dou, Y., & Zhao, X. (2023) · International Journal of Pattern Recognition and Artificial Intelligence · doi:10.1142/s0218001423570185
Smart Vision: Real Time Object DetectionPreprint
Narayanan, R. S., Abisharan, B., & Konetigari, L. R. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4636010
Classification and Localization of Objects Using Faster RCNN
Dakey, B. S., Ramani, G., Shabber, M., Jogu, S., Javvaji, D. S., & Jangam, H. (2024) · Lecture notes in networks and systems · doi:10.1007/978-981-99-8451-0_24
Vehicle Detection Based on Improved YOLOv5
Guo, H. (2024) · Advances in transdisciplinary engineering · doi:10.3233/atde240134
Chen, A., Wu, L., Winfree, S., Dunn, K. W., Salama, P., & Delp, E. J. (2024) · SPIE Medical Imaging: Image Processing · doi:10.1117/12.3005440
“To alleviate the effects bias 50 of selecting a certain IoU threshold to compute the evaluation metrics, we calculated the mean Precision, mean Recall and mean F 1 scores over multiple IoU thresholds.”
MethodResearch on Real-Time Pedestrian Detection Based on Infrared-Visible Image Fusion
Shang, Z., Cheng, B., Xie, X., Fu, T., & Wu, Z. (2023) · International Conference on Sensing, Measurement & Data Analytics in the era of Artificial Intelligence (ICSMD) · doi:10.1109/icsmd60522.2023.10490542
“Various strategies are employed for this purpose, including sliding-window methods [6], particle-window methods [7], [8], objectness methods [9], [10], and region proposal networks [11], [12].”
MethodDeléarde, R. (2022) · theses.fr (ABES) · doi:10.70675/f9cf337cz28d3z4ef3z8502z2215a1df6ab1
Iskandarova, S., Tulaganova, F., Mohidil, A., & Khairullo, K. (2024) · Международный Журнал Теоретических и Прикладных Вопросов Цифровых Технологий · doi:10.62132/ijdt.v7i2.177
“In [11,12] A threshold and mathematical morphology [13] were used to distinguish the cell nucleus.”
MethodLocalizing plant leaves using maximum anchor boxes in region proposal convolutional neural networks
Misra, D., Duraisamy, P., & Sandan, T. (2024) · SPIE Conference on Sensing for Agriculture and Food Quality and Safety · doi:10.1117/12.3015526
葛, 旭., 金, 学., 马, 慧., & 邹, 天. (2024) · Zhìnéng jīqìrén · doi:10.52810/jir.2024.004
Research on Robot Grasping Control Technology in Dense Stacking Environments Based on Deep Learning
Yang, L. (2024) · International Conference on Computing, Robotics and System Sciences (ICRSS) · doi:10.1109/icrss65752.2024.00023
A Scale Adaptive Tracking Algorithm Based on Feature Point Fitting
Yang, X., Zhang, L., & Sun, Y. (2025) · Lecture notes in electrical engineering · doi:10.1007/978-981-96-2204-7_2
Aligning Object Detector Bounding Boxes with Human Preference
Strafforello, O., Kayhan, O. S., Inel, O., Schutte, K., & van Gemert, J. C. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-92591-7_10 · Cited by 1
Holzapfel, K., Schöler, T., Dechent, R., Naumann, M. K., Allweyer, M., & Eisenacher, T. (2025) · Klinische Neurophysiologie · doi:10.1055/a-2573-0858
“Generell wird in der Literatur ein IoU > 0,5 als ausreichender Wert für eine Objekterkennung ange - sehen [20].”
Millimetre Wave Radar in Residual Neural Network with Multiple Large Model Parameters
Leong, W. Y., & Zhao, C. (2025) · ASM Science Journal · doi:10.32802/asmscj.2025.1955
Elangovan, R., Karthikeyan, G., Ramalingam, V., & Rajamanogaran, M. (2025) · International Conference on Advances in Modern Age Technologies for Health and Engineering Science (AMATHE) · doi:10.1109/amathe65477.2025.11081174
Learning-Based Distance Estimation for 360° Single-Sensor Setups
Quan, Y., Kiefer, B., Messmer, M., & Zell, A. (2025) · European Conference on Mobile Robots · doi:10.1109/ecmr65884.2025.11163323
“…performance in both object detection and distance estimation, we use mean Average Precision (mAP) [17] at different Intersection over Union (IoU) thresholds [18] for object detection, and assess distance estimation using both absolute and confidence-weighted error metrics, which defined as following.”
Improved YOLOv8-based video supervision technology for ulva prolifera salvage safety operations
Shang, H., Yang, K., Liu, Y., Ai, B., Liu, X., & Li, T. (2025) · International Conference on Electronics and Communication, Network and Computer Technology (ECNCT) · doi:10.1109/ecnct66493.2025.11172538
“Two-stage detectors such as R-CNN, Fast R-CNN and Faster R-CNN achieve high accuracy detection through candidate region generation and classification regression, but are computationally intensive and slow [4-6].”
Improve the Identification Accuracy of Parking Stickers with Persistent Homology Features
Zhang, X., Zhong, W., & Gu, F. (2025) · International Conference on Intelligent Computing and Robotics (ICICR) · doi:10.1109/icicr65456.2025.00148
“Besides the detection models, quality and size of a dataset play a crucial role in the detection performance as well [14].”
Region Enhanced Edge-Based Multi-Class Object Proposal for Self-Driving Vehicles
Haq, M. A., Huy, L. N. Q., & Ridlwan, M. (2025) · Khazanah Informatika Jurnal Ilmu Komputer dan Informatika · doi:10.23917/khif.v11i1.4662
“Additionally, the advancement of the classifier network has led to significant improvement of detection quality at the cost of increasingly complex and computationally expensive networks [9].”
InfluentialAirplane Detection Based on Unsupervised Deep Domain Adaptation in Remote Sensing ImagesPreprint
Youssef, Y. B., Lyaqini, S., Fakhar, K., & Abdelmounim, E. (2022) · Research Square · doi:10.21203/rs.3.rs-2088221/v1
Physical Reasoning for Intelligent Agent in Simulated Environments
Ge, X. (2017) · ANU Open Research (Australian National University) · doi:10.25911/5d723c78b18b0
Deep Neural Networks in Medical Microbiology - Bacterial Colonies Classification
Pereira, J. D. P. (2023) · Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT) · doi:10.34626/4ejb-yc24
System of Autonomous Aerial Vehicles for Subterranean Exploration with Slam Capabilities
Blanchard, N., & Blanchard, N. (2024) · UA Campus Repository (The University of Arizona)
Human pose and action recognition
Farrajota, M. (2017) · Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT)
Espinosa, J. E. (2019) · Repositorio Institucional UN - Biblioteca Digital · Cited by 2
Marciniak, K., Majewski, P., Lampa, P., Mrzygłód, M., & Reiner, J. (2025) · Advances in Science and Technology – Research Journal · doi:10.12913/22998624/209916
“Using the pycocotools library, average precision (AP50) [25] and average recall (AR50) [26] were determined.”
Li, Z. (2023) · Tuwhera (Auckland University of Technology)
Detecting people and content in educational videos
Pratusevich, M. (2015) · DSpace@MIT (Massachusetts Institute of Technology)
Towards robust visual object tracking : proposal selection and occlusion reasoning
Hua, Y. (2016) · theses.fr (ABES) · doi:10.70675/6198605ez66dfz4543zae20zb11f2812387e
“, 2015], including candidates from object proposals [Hosang et al., 2016] extracted for the entire frame, can improve tracking robustness, in particular for fast moving objects and low-frame-rate videos.”
BackgroundComputer Vision for Healthcare and Medicine: Unlocking Insights from Visual Data
Ray, A., Firouzi, F., Farahani, B., & Chakrabarty, K. (2026) · Springer eBooks · doi:10.1007/978-3-032-06286-4_5 · Cited by 1
New methods for image classification, image retrieval and semantic correspondence
de Rezende, R. S. (2017) · theses.fr (ABES) · doi:10.70675/0bd94776z371ez41e3z9afeze9c993ed9ab8
Video inpainting and semi-supervised object removal
Le, T. T. (2019) · theses.fr (ABES) · doi:10.70675/1cf4c899z91b9z479czafcdz6dce7a3b070a
“2017; N. Xu, Price, Cohen, J. Yang, et al. 2017). Several attempts have tried to combine the advantages of both object-proposals-based and object-detection-based approaches. For example, Zagoruyko et al. (2016) use a modified R-CNN model (Girshick et al. 2014) to propose instance bounding boxes as a starting point. Af…”
MethodHoang, K. (2021) · theses.fr (ABES) · doi:10.70675/e36ffbb4zbf05z4358za463zc57f0652a48e · Cited by 1
Semantic Analysis of the Driving Environment in Urban Scenarios
Lateef, F. (2021) · theses.fr (ABES) · doi:10.70675/114e131azb18az4b50zb9d8z55d904a50cbc
Deep neural networks in medical microbiology for bacterial colonies classification
Pereira, J. D., Veloso, B., & da Gama, J. M. P. (2026) · Scientific Reports · doi:10.1038/s41598-026-48621-0
“These metrics collectively provide a robust framework for assessing model performance, balancing precision, recall, and localization accuracy 30 .”
Competitive fusion in multimodal networks for enhanced salient object detection
Tan, H., Wen, S., Zhang, L., Li, J., & Hu, T. (2026) · The Visual Computer · doi:10.1007/s00371-026-04602-y · Cited by 1
“The predominant approach involves extracting RGB and depth features via multi-layer convolutional networks, which are then integrated using concatenation or attention mechanisms [23, 24].”
Region-based Network for Yoga Pose Estimation with Discriminative Fine-Tuning Optimization
Gite, S., Mane, D., Mane, V., Kale, S. D., & Dhotre, P. (2023) · International Journal of Intelligent Systems and Applications in Engineering
Chen, J., Liu, Z., Liu, R., Fang, Y., Zuo, Y., & Yan, J. (2026) · doi:10.1109/cvprf71909.2026.00756
“In the early days, low-level features ( e.g. , edges, super-pixels and saliency) are employed to produce bounding box proposals [1, 19, 73] or mask proposals [20, 22, 24, 25, 56].”
Fei, T., He, X., Zhang, B., & Wang, B. (2026) · Big Data and Cognitive Computing · doi:10.3390/bdcc10100355
Meta-Reinforcement Learning for Zero-Shot Adaptation in GNSS-Denied UAV Target Tracking
Almalis, N. D., Armenakis, N., & Tsihrintzis, G. A. (2026) · Procedia Computer Science · doi:10.1016/j.procs.2026.09.163
Xu, M., Gong, H., Hao, Z., Peng, J., Zheng, Y., & Yue, S. (2026) · Biomimetics · doi:10.3390/biomimetics11110787
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.