Multi-cue onboard pedestrian detection
The DOI matches its Crossref record; 36 of the 37 references it cites were found in a registry or an index.
Abstract
Various powerful people detection methods exist. Surprisingly, most approaches rely on static image features only despite the obvious potential of motion information for people detection. This paper systematically evaluates different features and classifiers in a sliding-window framework. First, our experiments indicate that incorporating motion information improves detection performance significantly. Second, the combination of multiple and complementary feature types can also help improve performance. And third, the choice of the classifier-feature combination and several implementation details are crucial to reach best performance. In contrast to many recent papers experimental results are reported for four different datasets rather than using a single one. Three of them are taken from the literature allowing for direct comparison. The fourth dataset is newly recorded using an onboard camera driving through urban environment. Consequently this dataset is more realistic and more challenging than any currently available dataset.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Citations
Cites
Histograms 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 5×
“Dalal et al. [6] enrich their static feature descriptor [ 5 ] with internal motion histograms to improve detection.”
BackgroundMethodResultInfluentialThe Ecological Approach to Visual Perception
Bornstein, M. H., & Gibson, J. J. (1980) · The Journal of Aesthetics and Art Criticism · doi:10.2307/429816 · Cited by 22,935 · cited 1×
“While psychologists and neuroscientists argue that motion is an important cue for human perception [17] only few computer vision object detectors (e.g. [30, 6]) exploit this fact.”
BackgroundRobust 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
Friedman, J., Hastie, T., & Tibshirani, R. (2000) · The Annals of Statistics · doi:10.1214/aos/1016218223 · Cited by 6,884 · cited 1×
“AdaBoost Contrary to SVMs, boosting algorithms [14] optimize the classification error on the training samples iteratively.”
MethodA discriminatively trained, multiscale, deformable part model
Felzenszwalb, P., McAllester, D., & Ramanan, D. (2008) · 2008 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2008.4587597 · Cited by 2,867 · cited 2×
“Popular classifiers are SVMs [26, 5, 13, 20, 19] or boosting [31, 7, 33, 23].”
BackgroundDetecting Pedestrians Using Patterns of Motion and Appearance
Viola, P., Jones, M. J., & Snow, D. (2005) · International Journal of Computer Vision · doi:10.1007/s11263-005-6644-8 · Cited by 2,079
A Duality Based Approach for Realtime TV-L 1 Optical Flow
Zach, C., Pock, T., & Bischof, H. (2007) · Lecture Notes in Computer Science · doi:10.1007/978-3-540-74936-3_22 · Cited by 1,658 · cited 2×
“de Objectives Detect pedestrians from a moving platform • Exploit motion information • Leverage complementarity of features • Evaluate different classifiers • New datasets with image pairs Features • HOG [1] 8× 8 pixel cells, 2× 2 blocks 9-bin histograms, unsigned gradients • Haar wavelets [2] 32 and 16 pixel masks ho…”
BackgroundHuman Detection Using Oriented Histograms of Flow and Appearance
Dalal, N., Triggs, B., & Schmid, C. (2006) · Lecture Notes in Computer Science · doi:10.1007/11744047_33 · Cited by 1,652 · cited 7×
“In contrast, [ 6 ] proposed motion features that are ‐ at least in principle ‐ applicable to onboard sequences.”
BackgroundMethodResultInfluentialPedestrian detection: A benchmark
Dollar, P., Wojek, C., Schiele, B., & Perona, P. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206631 · Cited by 1,313 · cited 3×
“Doll´ ar et al. [ 8 ] report results for various recent approaches on a new challenging onboard dataset with improved evaluation metrics.”
BackgroundInfluentialSelf-citationA Trainable System for Object Detection
Papageorgiou, C., & Poggio, T. (2000) · International Journal of Computer Vision · doi:10.1023/a:1008162616689 · Cited by 1,289 · cited 3×
“Haar Haar wavelets have been introduced by Papageor-giou&Poggio [22] for people detection.”
BackgroundMethodInfluentialMonocular Pedestrian Detection: Survey and Experiments
Enzweiler, M., & Gavrila, D. M. (2009) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2008.260 · Cited by 1,213 · cited 1×
“Enzweiler&Gavrila [9] evaluate different detectors in combination with an onboard system with focus on performance and runtime.”
BackgroundDetecting pedestrians using patterns of motion and appearance
Viola, Jones, & Snow (2003) · Proceedings Ninth IEEE International Conference on Computer Vision · doi:10.1109/iccv.2003.1238422 · Cited by 1,002
Pedestrian Detection via Classification on Riemannian Manifolds
Tuzel, O., Porikli, F., & Meer, P. (2008) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2008.75 · Cited by 975
Global data association for multi-object tracking using network flows
Zhang, L., Li, Y., & Nevatia, R. (2008) · 2008 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2008.4587584 · Cited by 925 · cited 1×
“We also outperform Zhang et al. [ 36 ] who report 64.3% recall at 1.5 FFPI even though their detector is trained on ETH-02 and ETH-03 whereas our detector is trained on an independent and more general multi-view training set.”
MethodInfluentialPeople-tracking-by-detection and people-detection-by-tracking
Andriluka, M., Roth, S., & Schiele, B. (2008) · 2008 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2008.4587583 · Cited by 824 · cited 2×
“Most detectors in this domain as well as ours employ the sliding-window scheme, but notable exceptions exist [21, 24, 1].”
BackgroundMethodSelf-citationA mobile vision system for robust multi-person tracking
Ess, A., Leibe, B., Schindler, K., & Van Gool, L. (2008) · 2008 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2008.4587581 · Cited by 721 · cited 2×
“While Ess et al. [11] extract the ground plane from a depth map and fuse it with detections in a graphical model, Ess et al. [12] add further 3D scene information by integration with Structure-from-Motion.”
BackgroundHuman Detection Based on a Probabilistic Assembly of Robust Part Detectors
Mikolajczyk, K., Schmid, C., & Zisserman, A. (2004) · Lecture Notes in Computer Science · doi:10.1007/978-3-540-24670-1_6 · Cited by 608
Depth and Appearance for Mobile Scene Analysis
Ess, A., Leibe, B., & Van Gool, L. (2007) · 2007 IEEE 11th International Conference on Computer Vision · doi:10.1109/iccv.2007.4409092 · Cited by 540
Multi-cue Pedestrian Detection and Tracking from a Moving Vehicle
Gavrila, D. M., & Munder, S. (2007) · International Journal of Computer Vision · doi:10.1007/s11263-006-9038-7 · Cited by 524 · cited 2×
“While [29] convincingly showed that multiple features improve performance for image classification, for detection only few approaches exploit this fact [34, 16, 32].”
BackgroundMethodLearning The Discriminative Power-Invariance Trade-Off
Varma, M., & Ray, D. (2007) · 2007 IEEE 11th International Conference on Computer Vision · doi:10.1109/iccv.2007.4408875 · Cited by 520 · cited 2×
“More sophisticated feature combination schemes (e.g. [29]) may allow to improve performance more consistently based on multiple features.”
BackgroundDetecting Pedestrians by Learning Shapelet Features
Sabzmeydani, P., & Mori, G. (2007) · 2007 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2007.383134 · Cited by 515 · cited 2×
“Popular classifiers are SVMs [26, 5, 13, 20, 19] or boosting [31, 7, 33, 23].”
BackgroundShashua, A., Gdalyahu, Y., & Hayun, G. (2004) · IEEE Intelligent Vehicles Symposium, 2004 · doi:10.1109/ivs.2004.1336346 · Cited by 352 · cited 1×
“Sashua et al. [ 26 ] use parts and employ histograms of gradients as features.”
MethodAn algorithm for data-driven bandwidth selection
Comaniciu, D. (2003) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2003.1177159 · Cited by 336 · cited 1×
“Here, we adopt an adapted bandwidth mean-shift based mode seeking strategy [3] to determine the position in x - y - scale -space, but determine the final detection’s score to be the maximum of all scores within the mode.”
MethodFinding People in Images and Videos
Dalal, N., Imagerie, S. ́., Robotique, V. E., Président, M. J. L. C., Martial, M., Luc, M., Shai, M., Examinateur, A., Cordelia, M., & Directeur, S. (2006) · OpenGrey (Institut de l'Information Scientifique et Technique) · Cited by 305
A Bayesian, Exemplar-Based Approach to Hierarchical Shape Matching
Gavrila, D. M. (2007) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2007.1062 · Cited by 268 · cited 1×
“Gavrila [15] proposes a tree-structured Bayesian approach that builds on offline clustering of pedestrian shapes.”
BackgroundA Performance Evaluation of Single and Multi-feature People Detection
Wojek, C., & Schiele, B. (2008) · Lecture Notes in Computer Science · doi:10.1007/978-3-540-69321-5_9 · Cited by 246 · cited 3×
“For people detection Wojek&Schiele [32] report results for various recent approaches on a new challenging onboard dataset with improved evaluation metrics.”
BackgroundSelf-citationCluster Boosted Tree Classifier for Multi-View, Multi-Pose Object Detection
Wu, B., & Nevatia, R. (2007) · 2007 IEEE 11th International Conference on Computer Vision · doi:10.1109/iccv.2007.4409006 · Cited by 188 · cited 2×
“In [ 33 ] Wu&Nevatia propose a system to automatically construct tree hierarchies for the problem of multi-view pedestrian detection.”
BackgroundA Pose-Invariant Descriptor for Human Detection and Segmentation
Lin, Z., & Davis, L. S. (2008) · Lecture notes in computer science · doi:10.1007/978-3-540-88693-8_31 · Cited by 167 · cited 2×
“Popular classifiers are SVMs [26, 5, 13, 20, 19] or boosting [31, 7, 33, 23].”
BackgroundMultiple Component Learning for Object Detection
Dollár, P., Babenko, B., Belongie, S., Perona, P., & Tu, Z. (2008) · Lecture notes in computer science · doi:10.1007/978-3-540-88688-4_16 · Cited by 153 · cited 1×
“Similarly, Doll´ ar et al. [ 7 ] present an approach that automatically learns flexible parts from training data and uses a boosting framework with wavelet features.”
BackgroundWu, B., & Nevatia, R. (2008) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2008.4587749 · Cited by 145 · cited 2×
“While [29] convincingly showed that multiple features improve performance for image classification, for detection only few approaches exploit this fact [ 34 , 16, 32].”
BackgroundSimultaneous Learning and Alignment: Multi-Instance and Multi-Pose Learning ?
Babenko, B., Tu, Z., & Belongie, S. (2008) · HAL · Cited by 86
Multi-Aspect Detection of Articulated Objects
Seemann, E., Leibe, B., & Schiele, B. (2006) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2006.193 · Cited by 80 · cited 1×
“Wu&Nevatia [33] remedy this issue by learning a tree structured classifier, Lin&Davis [19] use a handcrafted hierarchy, while Seemann et al. [25] propose multi-articulation learning.”
BackgroundSelf-citationTowards Robust Pedestrian Detection in Crowded Image Sequences
Seemann, E., Fritz, M., & Schiele, B. (2007) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2007.383300 · Cited by 62 · cited 1×
“Most detectors in this domain as well as ours employ the sliding-window scheme, but notable exceptions exist [21, 24, 1].”
MethodSelf-citationConfiguration Estimates Improve Pedestrian Finding
Tran, D., & Forsyth, D. A. (2007) · Illinois Digital Environment for Access to Learning and Scholarship (University of Illinois at Urbana-Champaign) · Cited by 47
Monocular pedestrian recognition using motion parallax
Enzweiler, M., Kanter, P., & Gavrila, D. M. (2008) · IEEE Intelligent Vehicles Symposium · doi:10.1109/ivs.2008.4621169 · Cited by 47 · cited 1×
“Enzweiler et al. [10] use motion information in a ROI generation step for an onboard system, while we investigate the use of motion features for the detector.”
MethodMCBoost: Multiple Classifier Boosting for Perceptual Co-clustering of Images and Visual Features
Kim, T., Cipolla, R., et al. (2009) · Cambridge University Engineering Department Publications Database · Cited by 45
Research works printed but not linked: 1
- maji, 2008, Classification using intersection kernel SVMs is efficient, CVPR
Cited by
nuScenes: A Multimodal Dataset for Autonomous Driving
Caesar, H., Bankiti, V., Lang, A. H., Vora, S., Liong, V. E., Xu, Q., Krishnan, A., Pan, Y., Baldan, G., & Beijbom, O. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.01164 · Cited by 7,152
“Other datasets focus exclusively on pedestrian annotations on images [20, 25, 79, 24, 88, 23, 58].”
BackgroundScalable Person Re-identification: A Benchmark
Zheng, L., Shen, L., Tian, L., Wang, S., Wang, J., & Tian, Q. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.133 · Cited by 4,506
“For other datasets, the codebook is trained on the independent TUD-Brussels dataset [35].”
MethodPedestrian Detection: An Evaluation of the State of the Art
Dollar, P., Wojek, C., Schiele, B., & Perona, P. (2012) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2011.155 · Cited by 3,190
“Pedestrians can be labeled in photographs [7], [16], surveillance video [17], [24], and images taken from a mobile recording setup, such as a robot or vehicle [4], [5], [6].”
BackgroundMethodInfluentialSelf-citationFast Feature Pyramids for Object Detection
Dollar, P., Appel, R., Belongie, S., & Perona, P. (2014) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2014.2300479 · Cited by 1,955
“We show results for both pedestrian detection (measured on the Caltech [31], INRIA [21], TUD- Brussels [36] and ETH [37] data sets) and general object detection (measured on the PASCAL VOC [38]).”
MethodA Survey of Deep Learning-Based Object Detection
Jiao, L., Zhang, F., Liu, F., Yang, S., Li, L., Feng, Z., & Qu, R. (2019) · IEEE Access · doi:10.1109/access.2019.2939201 · Cited by 1,323
“TUD-Brussels[57] mobile 1776 218 1092 1498 508”
Metadata
- Authors
- Wojek, C., Walk, S., & Schiele, B.
- Affiliations
- Technische Universität Darmstadt (DE)
- ORCID
- Schiele, Bernt
- Venue
- 2009 IEEE Conference on Computer Vision and Pattern Recognition
- Publisher
- IEEE
- Published
- 2009-06
- Pages
- 794-801
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W2127420331
- Cited by
- 350
- References
- 37
- Citations by year
- 2026: 2 · 2025: 5 · 2024: 1 · 2023: 8 · 2022: 13 · 2021: 21 · 2020: 24 · 2019: 25 · 2018: 27 · 2017: 30 · 2016: 24 · 2015: 30 · 2014: 29 · 2013: 37 · 2012: 19
- FWCI
- 20.14
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- Max Planck Digital Library (submittedVersion)
- MPG.PuRe (Max Planck Society) (submittedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Video Surveillance and Tracking Methods
- Physical Sciences › Engineering › Automotive Engineering › Autonomous Vehicle Technology and Safety
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Keywords
- pedestrian detection, onboard camera, motion information, multi-cue fusion, feature combination, sliding window, classifier comparison, urban driving
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.