Academia.eduAcademia.edu

Outline

Title
Abstract
Introduction
Related Work
Methodology
Results and Analysis
Conclusion
References
All Topics
Computer Science
Computer Vision

The Impact of Partial Occlusion on Pedestrian Detectability

Profile image of Martin GlavinMartin Glavin

2022, arXiv (Cornell University)

https://doi.org/10.48550/ARXIV.2205.04812
visibility

…

description

9 pages

link

1 file

Abstract

Robust detection of vulnerable road users is a safety critical requirement for the deployment of autonomous vehicles in heterogeneous traffic. One of the most complex outstanding challenges is that of partial occlusion where a target object is only partially available to the sensor due to obstruction by another foreground object. A number of leading pedestrian detection benchmarks provide annotation for partial occlusion, however each benchmark varies greatly in their definition of the occurrence and severity of occlusion. Recent research demonstrates that a high degree of subjectivity is used to classify occlusion level in these cases and occlusion is typically categorized into 2-3 broad categories such as "partially" and "heavily" occluded. This can lead to inaccurate or inconsistent reporting of pedestrian detection model performance depending on which benchmark is used. This research introduces a novel, objective benchmark for partially occluded pedestrian detection to facilitate the objective characterization of pedestrian detection models. Characterization is carried out on seven popular pedestrian detection models for a range of occlusion levels from 0-99%. Results demonstrate that pedestrian detection performance degrades, and the number of false negative detections increase as pedestrian occlusion level increases. Of the seven popular pedestrian detection routines characterized, CenterNet has the greatest overall performance, followed by SSDlite. Reti-naNet has the lowest overall detection performance across the range of occlusion levels.

Related papers

Pedestrian Detection in Automotive Safety: Understanding State-of-the-Art
Rajesh R, Ragesh NK

Pedestrian detection is one of the important computer vision problems in automotive safety and driver assistance domain. It is a major component of the advanced driver assistance system (ADAS) which help the driver to drive safely. Recent literature shows a number of research activities addressing object detection/tracking in general and pedestrian detection in particular. The solutions proposed by different researchers vary in detection methods, detection scenario, feature descriptors, classification schemes, detection performance, as well as computational complexity. However, the average detection accuracy is not much promising even after many years of research. The fail-safe and real-time human detection from real life road scenes, even in standard resolution, is far from reality. Safety critical systems in the automotive industry have to follow well established stringent safety standards like ISO26262. Since the pedestrian detection system deals with human safety, it also has to follow these standards before integrating to the vehicle electronics. This paper is a study of different techniques used in pedestrian detection specific to the automotive application, along with a description of generic pedestrian detection solution architecture. INDEX TERMS Road safety, advanced driver assistance systems (ADAS), smart image region of interest (ROI) selection, image feature descriptors, pedestrian detection.

View PDFchevron_right
A Robust Pedestrian Detection Approach for Autonomous Vehicles
Ali Tourani

2022 8th Iranian Conference on Signal Processing and Intelligent Systems (ICSPIS)

Nowadays, utilizing Advanced Driver-Assistance Systems (ADAS) has absorbed a huge interest as a potential solution for reducing road traffic issues. Despite recent technological advances in such systems, there are still many inquiries that need to be overcome. For instance, ADAS requires accurate and real-time detection of pedestrians in various driving scenarios. To solve the mentioned problem, this paper aims to fine-tune the YOLOv5s framework for handling pedestrian detection challenges on the real-world instances of Caltech pedestrian dataset. We also introduce a developed toolbox for preparing training and test data and annotations of Caltech pedestrian dataset into the format recognizable by YOLOv5. Experimental results of utilizing our approach show that the mean Average Precision (mAP) of our fine-tuned model for pedestrian detection task is more than 91 percent when performing at the highest rate of 70 FPS. Moreover, the experiments on the Caltech pedestrian dataset samples have verified that our proposed approach is an effective and accurate method for pedestrian detection and can outperform other existing methodologies.

View PDFchevron_right
Part-Based Pedestrian Detection and Feature-Based Tracking for Driver Assistance: Real-Time, Robust Algorithms, and Evaluation
Mohan Trivedi

IEEE Transactions on Intelligent Transportation Systems, 2000

Detecting pedestrians is still a challenging task for automotive vision systems due to the extreme variability of targets, lighting conditions, occlusion, and high-speed vehicle motion. Much research has been focused on this problem in the last ten years and detectors based on classifiers have gained a special place among the different approaches presented. This paper presents a state-of-the-art pedestrian detection system based on a two-stage classifier. Candidates are extracted with a Haar cascade classifier trained with the Daimler Detection Benchmark data set and then validated through a part-based histogram-of-orientedgradient (HOG) classifier with the aim of lowering the number of false positives. The surviving candidates are then filtered with feature-based tracking to enhance the recognition robustness and improve the results' stability. The system has been implemented on a prototype vehicle and offers high performance in terms of several metrics, such as detection rate, false positives per hour, and frame rate. The novelty of this system relies on the combination of a HOG part-based approach, tracking based on a specific optimized feature, and porting on a real prototype. Index Terms-Advanced driver assistance system (ADAS), classifiers, features, machine vision, pedestrian detection. I. INTRODUCTION O VER the past decade, the essential role of machine vision modules to realize active safety systems for accident prevention is clearly established in academic research [1], [2] and is also reflected in innovative systems introduced by industry [3], [4]. Effective vision systems need to accurately assess situational criticalities from the panoramic surround of a vehicle [5] and simultaneously assess awareness of these criticalities by the driver [6]. One of the major thrusts in situational criticality assessment is that of pedestrian detection, and Manuscript

View PDFchevron_right
Pedestrian and Vehicle Detection in Autonomous Vehicle Perception Systems—A Review
Tatiana Kalganova

Sensors

Autonomous Vehicles (AVs) have the potential to solve many traffic problems, such as accidents, congestion and pollution. However, there are still challenges to overcome, for instance, AVs need to accurately perceive their environment to safely navigate in busy urban scenarios. The aim of this paper is to review recent articles on computer vision techniques that can be used to build an AV perception system. AV perception systems need to accurately detect non-static objects and predict their behaviour, as well as to detect static objects and recognise the information they are providing. This paper, in particular, focuses on the computer vision techniques used to detect pedestrians and vehicles. There have been many papers and reviews on pedestrians and vehicles detection so far. However, most of the past papers only reviewed pedestrian or vehicle detection separately. This review aims to present an overview of the AV systems in general, and then review and investigate several detecti...

View PDFchevron_right
PSC-Net: learning part spatial co-occurrence for occluded pedestrian detection
rao anwer

Science China Information Sciences, 2020

Detecting pedestrians, especially under heavy occlusion, is a challenging computer vision problem with numerous real-world applications. This paper introduces a novel approach, termed as PSC-Net, for occluded pedestrian detection. The proposed PSC-Net contains a dedicated module that is designed to explicitly capture both inter and intra-part co-occurrence information of different pedestrian body parts through a graph convolutional network (GCN). Both inter and intra-part co-occurrence information contribute towards improving the feature representation for handling varying level of occlusions, ranging from partial to severe occlusions. Our PSC-Net exploits the topological structure of pedestrian and does not require part-based annotations or additional visible bounding-box (VBB) information to learn part spatial co-occurrence. Comprehensive experiments are performed on three challenging datasets: CityPersons, Caltech, and CrowdHuman datasets. Particularly, in terms of log-average miss rates and with the same backbone and input scale as those of the state-of-the-art MGAN, the proposed PSC-Net achieves absolute gains of 4.0% and 3.4% over MGAN on the heavy occlusion subsets of CityPersons and Caltech test sets, respectively.

View PDFchevron_right
Survey of pedestrian detection for advanced driver assistance systems
Antonio Peña

Pattern Analysis and Machine Intelligence, IEEE Transactions on, 2010

Advanced driver assistance systems (ADASs), and particularly pedestrian protection systems (PPSs), have become an active research area aimed at improving traffic safety. The major challenge of PPSs is the development of reliable on-board pedestrian detection systems. Due to the varying appearance of pedestrians (eg, different clothes, changing size, aspect ratio, and dynamic shape) and the unstructured environment, it is very difficult to cope with the demanded robustness of this kind of system.

View PDFchevron_right
Bounding Box Accuracy in Pedestrian Detection for Intelligent Transportation Systems
David Fernandez

2006

This paper describes a stereo-vision-based pedestrian detection system for intelligent transportation systems. The basic components of pedestrians are first located in the image and then combined with a SVM-based classifier. Generic obstacles are located using a subtractive clustering attention mechanism based on stereo vision. A by-components learning approach is proposed and different feature extraction methods are tested in order to better deal with pedestrian variability and justify what features are better to be learnt for pedestrian detection. Candidate selection mechanisms usually yield pedestrians with inaccurate bounding boxes. Then a decrease in detection rate takes place if the SVM classifier is trained only with well-fitted pedestrians. Using several off-line databases containing thousands of pedestrians samples the effect of bounding box accuracy is studied. A multi-candidate generation mechanism is also developed in order to enhance the single frame performance, decreasing the number of false positives due to inaccurate bounding boxes

View PDFchevron_right
Improving the Pedestrian Detection Performance in the Absence of Rich Training Datasets: A UK Case Study
Tabassom Sedighi

Advances in Artificial Intelligence and Machine Learning

The World Health Organization estimates that well in excess of one million of lives are lost each year due to road traffic accidents. Since the human factor is the preeminent cause behind the traffic accidents, the development of reliable Advanced Driver Assistance Systems (ADASs) and Autonomous Vehicles (AVs) is seen by many as a possible solution to improve road safety. ADASs rely on the car perception system input that consists of camera(s), LIDAR and/or radar to detect pedestrians and other objects on the road. Hardware improvements as well as advances done in employing Deep Learning techniques for object detection popularized the Convolutional Neural Networks in the area of autonomous driving research and applications. However, the availability of quality and large datasets continues to be a most important contributor to the Deep Learning based model’s performance. With this in mind, this work analyses how a YOLO-based object detection architecture responded to limited data ava...

View PDFchevron_right
V2F-Net: Explicit Decomposition of Occluded Pedestrian Detection
Dawei Xiang

ArXiv, 2021

Occlusion is very challenging in pedestrian detection. In this paper, we propose a simple yet effective method named V2F-Net, which explicitly decomposes occluded pedestrian detection into visible region detection and full body estimation. V2F-Net consists of two sub-networks: Visible region Detection Network (VDN) and Full body Estimation Network (FEN). VDN tries to localize visible regions and FEN estimates full-body box on the basis of the visible box. Moreover, to further improve the estimation of full body, we propose a novel Embedding-based Part-aware Module (EPM). By supervising the visibility for each part, the network is encouraged to extract features with essential part information. We experimentally show the effectiveness of V2F-Net by conducting several experiments on two challenging datasets. V2F-Net achieves 5.85% AP gains on CrowdHuman and 2.24% MR−2 improvements on CityPersons compared to FPN baseline. Besides, the consistent gain on both one-stage and two-stage dete...

View PDFchevron_right
Stereo-Based Framework for Pedestrian Detection with Partial Occlusion Handling
Vittorio Murino

2012 IEEE Ninth International Conference on Advanced Video and Signal-Based Surveillance, 2012

The pedestrian detection literature has been recently renewed by the availability of large-scale multisensory datasets, able to capture complementary aspects of the objects of interest, namely, appearance, motion, and depth. In this paper, we exploit this multimodal scenario to propose a new set of composite descriptors dubbed CO 2 , COvariances of visual features and CO-occurrences of depth fields. Covariances of visual features allow to integrate at low-level heterogeneous visual cues related to intensity and texture. Co-occurrences of depth fields are brand new descriptors, which use range information for characterizing the global shape of a pedestrian while being also able to identify its occluded parts. This paper illustrates how these descriptors can be instantiated and combined together for improving the detection capabilities, just taking benefit from the proper handling of occlusions. Experimental results show that CO 2 , fed into a standard discriminative classification system, allow to set state-of-the-art performances on recent multimodal intensity-and stereo-based pedestrian datasets.

View PDFchevron_right

References (44)

  1. S. international, "Taxonomy and definitions for terms related to driving automation systems for on-road motor vehicles," SAE, 2018. 1
  2. S. Gilroy, E. Jones, and M. Glavin, "Overcoming occlusion in the automotive environment-a review," IEEE Transactions on Intelligent Transportation Systems, 2019. 1, 2
  3. C. Ning, L. Menglu, Y. Hao, S. Xueping, and L. Yunhong, "Survey of pedestrian detection with occlusion," Complex & Intelligent Systems, vol. 7, no. 1, pp. 577-587, 2021. 1, 2
  4. J. Cao, Y. Pang, J. Xie, F. S. Khan, and L. Shao, "From hand- crafted to deep features for pedestrian detection: A survey," IEEE Transactions on Pattern Analysis and Machine Intelli- gence, 2021. 1
  5. Y. Xiao, K. Zhou, G. Cui, L. Jia, Z. Fang, X. Yang, and Q. Xia, "Deep learning for occluded and multi-scale pedes- trian detection: A review," IET Image Processing, 2021. 1, 2
  6. S. Gilroy, M. Glavin, E. Jones, and D. Mullins, "Pedes- trian occlusion level classification using keypoint detection and 2d body surface area estimation," in Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 3833-3839, 2021. 1, 2
  7. A. Geiger, P. Lenz, and R. Urtasun, "Are we ready for au- tonomous driving? the kitti vision benchmark suite," in 2012 IEEE Conference on Computer Vision and Pattern Recogni- tion, pp. 3354-3361, IEEE, 2012. 1, 2
  8. S. Zhang, R. Benenson, and B. Schiele, "Citypersons: A di- verse dataset for pedestrian detection," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recogni- tion, pp. 3213-3221, 2017. 1, 2
  9. M. Braun, S. Krebs, F. Flohr, and D. M. Gavrila, "Eurocity persons: A novel benchmark for person detection in traffic scenes," IEEE transactions on pattern analysis and machine intelligence, vol. 41, no. 8, pp. 1844-1861, 2019. 1, 2
  10. S. Ren, K. He, R. Girshick, and J. Sun, "Faster r-cnn: to- wards real-time object detection with region proposal net- works," IEEE transactions on pattern analysis and machine intelligence, vol. 39, no. 6, pp. 1137-1149, 2016. 3, 5
  11. B. E. Moore and J. J. Corso, "Fiftyone model zoo," Online. https://voxel51.com/docs/fiftyone/user guide/model zoo, 2022. 3
  12. K. He, G. Gkioxari, P. Dollár, and R. Girshick, "Mask r- cnn," in Proceedings of the IEEE international conference on computer vision, pp. 2961-2969, 2017. 3, 5
  13. J. Dai, Y. Li, K. He, and J. Sun, "R-fcn: Object detection via region-based fully convolutional networks," in Advances in neural information processing systems, pp. 379-387, 2016. 3, 5
  14. W. Liu, D. Anguelov, D. Erhan, C. Szegedy, S. Reed, C.- Y. Fu, and A. C. Berg, "Ssd: Single shot multibox detec- tor," in European conference on computer vision, pp. 21-37, Springer, 2016. 3, 4, 5
  15. Pytorch, "Torchvision model zoo," Online. https://pytorch.org/vision/stable/models.html, 2022. 3
  16. A. Howard, M. Sandler, G. Chu, L.-C. Chen, B. Chen, M. Tan, W. Wang, Y. Zhu, R. Pang, V. Vasudevan, et al., "Searching for mobilenetv3," in Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 1314-1324, 2019. 3, 4, 5
  17. M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, "Mobilenetv2: Inverted residuals and linear bottle- necks," in Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 4510-4520, 2018. 3, 4, 5
  18. T.-Y. Lin, P. Goyal, R. Girshick, K. He, and P. Dollár, "Focal loss for dense object detection," in Proceedings of the IEEE international conference on computer vision, pp. 2980- 2988, 2017. 3, 4, 5
  19. X. Zhou, D. Wang, and P. Krähenbühl, "Objects as points," arXiv preprint arXiv:1904.07850, 2019. 3, 4, 5
  20. P. Dollar, C. Wojek, B. Schiele, and P. Perona, "Pedestrian detection: An evaluation of the state of the art," IEEE trans- actions on pattern analysis and machine intelligence, vol. 34, no. 4, pp. 743-761, 2011. 2
  21. P. Dollár, C. Wojek, B. Schiele, and P. Perona, "Pedestrian detection: A benchmark," in 2009 IEEE Conference on Com- puter Vision and Pattern Recognition, pp. 304-311, IEEE, 2009. 2
  22. S. Zhang, R. Benenson, M. Omran, J. Hosang, and B. Schiele, "How far are we from solving pedestrian detec- tion?," in Proceedings of the iEEE conference on computer vision and pattern recognition, pp. 1259-1267, 2016. 2
  23. Y. Pang, J. Cao, Y. Li, J. Xie, H. Sun, and J. Gong, "Tju- dhd: A diverse high-resolution dataset for object detection," IEEE Transactions on Image Processing, vol. 30, pp. 207- 219, 2020. 2
  24. S. Shao, Z. Zhao, B. Li, T. Xiao, G. Yu, X. Zhang, and J. Sun, "Crowdhuman: A benchmark for detecting human in a crowd," arXiv preprint arXiv:1805.00123, 2018. 2
  25. C. Chi, S. Zhang, J. Xing, Z. Lei, S. Z. Li, and X. Zou, "Pedhunter: Occlusion robust pedestrian detector in crowded scenes," in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 34, pp. 10639-10646, 2020. 2
  26. X. Li, L. Li, F. Flohr, J. Wang, H. Xiong, M. Bernhard, S. Pan, D. M. Gavrila, and K. Li, "A unified framework for concurrent pedestrian and cyclist detection," IEEE transac- tions on intelligent transportation systems, vol. 18, no. 2, pp. 269-281, 2016. 2
  27. S. Hwang, J. Park, N. Kim, Y. Choi, and I. So Kweon, "Mul- tispectral pedestrian detection: Benchmark dataset and base- line," in Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 1037-1045, 2015. 2
  28. S. Walk, N. Majer, K. Schindler, and B. Schiele, "New fea- tures and insights for pedestrian detection," in 2010 IEEE Computer society conference on computer vision and pattern recognition, pp. 1030-1037, IEEE, 2010. 2
  29. R. N. Rajaram, E. Ohn-Bar, and M. M. Trivedi, "An explo- ration of why and when pedestrian detection fails," in 2015 IEEE 18th International Conference on Intelligent Trans- portation Systems, pp. 2335-2340, IEEE, 2015. 2
  30. J. Mao, T. Xiao, Y. Jiang, and Z. Cao, "What can help pedes- trian detection?," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3127-3136, 2017. 2
  31. S. Zhang, R. Benenson, M. Omran, J. Hosang, and B. Schiele, "Towards reaching human performance in pedes- trian detection," IEEE transactions on pattern analysis and machine intelligence, vol. 40, no. 4, pp. 973-986, 2017. 2
  32. N. Ragesh and R. Rajesh, "Pedestrian detection in automo- tive safety: understanding state-of-the-art," IEEE Access, vol. 7, pp. 47864-47890, 2019. 2
  33. J. Cao, Y. Pang, J. Han, B. Gao, and X. Li, "Taking a look at small-scale pedestrians and occluded pedestrians," IEEE transactions on image processing, vol. 29, pp. 3143-3152, 2019. 2
  34. T. Toprak, B. A. Can, M. Ozcelikors, S. B. Tekin, and M. A. Selver, "Limitations of feature-classifier strategies on pedes- trian detection for self driving cars," in 2020 International Conference on Electrical, Communication, and Computer Engineering (ICECCE), pp. 1-6, IEEE, 2020. 2
  35. I. Hasan, S. Liao, J. Li, S. U. Akram, and L. Shao, "Gen- eralizable pedestrian detection: The elephant in the room," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 11328-11337, 2021. 2
  36. H. Chandel and S. Vatta, "Occlusion detection and handling: a review," International Journal of Computer Applications, vol. 120, no. 10, 2015. 2
  37. K. Saleh, S. Szénási, and Z. Vámossy, "Occlusion handling in generic object detection: A review," in 2021 IEEE 19th World Symposium on Applied Machine Intelligence and In- formatics (SAMI), pp. 000477-000484, IEEE, 2021. 2
  38. S. Gilroy, M. Glavin, E. Jones, and D. Mullins, "An objective method for pedestrian occlusion level classification," arXiv preprint arXiv:2205.05412, 2022. 2
  39. A. Wallace, "The exposure treatment of burns," The Lancet, vol. 257, no. 6653, pp. 501-504, 1951. 2
  40. W.-S. Zheng, X. Li, T. Xiang, S. Liao, J. Lai, and S. Gong, "Partial person re-identification," in Proceedings of the IEEE International Conference on Computer Vision, pp. 4678- 4686, 2015. 2
  41. J. Zhuo, Z. Chen, J. Lai, and G. Wang, "Occluded person re-identification," in 2018 IEEE International Conference on Multimedia and Expo (ICME), pp. 1-6, IEEE, 2018. 2
  42. T.-Y. Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ra- manan, P. Dollár, and C. L. Zitnick, "Microsoft coco: Com- mon objects in context," in European conference on com- puter vision, pp. 740-755, Springer, 2014. 3
  43. L. Chen, S. Lin, X. Lu, D. Cao, H. Wu, C. Guo, C. Liu, and F.-Y. Wang, "Deep neural network based vehicle and pedes- trian detection for autonomous driving: A survey," IEEE Transactions on Intelligent Transportation Systems, 2021. 3
  44. B. E. Moore and J. J. Corso, "Fiftyone," GitHub. Note: https://github.com/voxel51/fiftyone, 2020. 4

Related papers

Occlusion-robust Detector Trained with Occluded Pedestrians
Wenzhi Liao

Proceedings of the 7th International Conference on Pattern Recognition Applications and Methods

Pedestrian detection has achieved a remarkable progress in recent years, but challenges remain especially when occlusion happens. Intuitively, occluded pedestrian samples contain some characteristic occlusion appearance features that can help to improve detection. However, we have observed that most existing approaches intentionally avoid using samples of occluded pedestrians during the training stage. This is because such samples will introduce unreliable information, which affects the learning of model parameters and thus results in dramatic performance decline. In this paper, we propose a new framework for pedestrian detection. The proposed method exploits the use of occluded pedestrian samples to learn more robust features for discriminating pedestrians, and enables better performances on pedestrian detection, especially for the occluded pedestrians (which always happens in many real applications). Compared to some recent detectors on Caltech Pedestrian dataset, with our proposed method, detection miss rate for occluded pedestrians are significantly reduced.

View PDFchevron_right
A cascade framework for unoccluded and occluded pedestrian detection
Irfan Ahmad

Proceedings of the 2014 IEEE Students' Technology Symposium, 2014

This paper presents a novel pedestrian detection framework for efficient detection of both unoccluded and occluded pedestrians, thereby proposing an efficient technique for pedestrian detection in real-time environment. Our framework consists of two layers of detection, the first layer using full body detectors for accurate detection of unoccluded pedestrians and then a cascaded layer of part based detectors to efficiently detect the occluded pedestrians. The full body detectors based techniques are state-of-the art for unoccluded pedestrian detection and the part based model is a viable choice for partially occluded pedestrian detection. In our part based model, we use six parts; three horizontal parts and three vertical parts thereby creating a model that is robust to varying degrees and types of occlusions. Each detection layer utilizes multiple modalities (cues) namely; intensity, dense stereo and dense flow. The use of part based detectors as the cascaded layer also increases the unoccluded pedestrian detection rate by correctly detecting the pedestrians that had been misclassified by the first layer. Thus, the second layer of part based detectors has a synergic effect on the first layer.

View PDFchevron_right
Pedestrian Detection: The Elephant In The Room
saad akram

2020

Pedestrian detection is used in many vision based applications ranging from video surveillance to autonomous driving. Despite achieving high performance, it is still largely unknown how well existing detectors generalize to unseen data. To this end, we conduct a comprehensive study in this paper, using a general principle of direct cross-dataset evaluation. Through this study, we find that existing state-of-the-art pedestrian detectors generalize poorly from one dataset to another. We demonstrate that there are two reasons for this trend. Firstly, they over-fit on popular datasets in a traditional single-dataset training and test pipeline. Secondly, the training source is generally not dense in pedestrians and diverse in scenarios. Accordingly, through experiments we find that a general purpose object detector works better in direct cross-dataset evaluation compared with state-of-the-art pedestrian detectors and we illustrate that diverse and dense datasets, collected by crawling th...

View PDFchevron_right
Towards Practical Evaluation of Pedestrian Detectors
Mohamed Hussein

2008

Despite recent significant advancement in the area of pedestrian detection in images, little effort has been devoted to algorithm evaluation for practical purposes. Typically, detectors are evaluated only on color images. It is not clear how the performance would be affected if other modalities are used, e.g. thermal or near infrared. Also, detectors are evaluated on cropped images that have the same size as training images. However, in practice, detectors are applied to large images with multiple pedestrians in different locations and sizes. To apply a single size pedestrian detector, the input image is, typically, scanned several times with different window sizes. It is not clear how the detection performance would be affected by such multiple-size scanning technique. Moreover, to implement such a technique, one is faced with a multitude of design choices, each of which potentially affects the performance of the detector. The contribution of this paper is to assess and reason about the differences in detection performance of two state of the art detectors across changes in modality (visible or near infrared), evaluation method (on cropped or whole images), the effect of different design choices (resizing features or images, smoothing or not).

View PDFchevron_right
Safety Analysis of Deep Learning based 2D Pedestrian Detectors in the Context of Autonomous Driving in Urban Traffic
Alen Smajic

2022

AI-based computer vision systems play a crucial role in the environment perception for autonomous driving. Although the development of self-driving systems has been pursued for multiple decades, it is only recently that breakthroughs in Deep Neural Networks (DNNs) have led to their widespread application in perception pipelines, which are getting more and more sophisticated. However, with this rising trend comes the need for a systematic safety analysis to evaluate the DNN's behavior in difficult scenarios as well as to identify the various factors that cause misbehavior in such systems. This work aims to deliver a crucial contribution to the lacking literature on the systematic analysis of Performance Limiting Factors (PLFs) for DNNs by investigating the task of pedestrian detection in urban traffic from a monocular camera mounted on an autonomous vehicle. To investigate the common factors that lead to DNN misbehavior, six commonly used state-of-the-art object detection architectures and three detection tasks are studied using a new large-scale synthetic dataset and a smaller real-world dataset for pedestrian detection. The systematic analysis includes 17 factors from the literature and four novel factors that are introduced as part of this work. Each of the 21 factors is assessed based on its influence on the detection performance and whether it can be considered a Performance Limiting Factor (PLF). In order to support the evaluation of the detection performance, a novel and task-oriented Pedestrian Detection Safety Metric (PDSM) is introduced, which is specifically designed to aid in the identification of individual factors that contribute to DNN failure. This work further introduces a training approach for F1-Score maximization whose purpose is to ensure that the DNNs are assessed at their highest performance. Moreover, a new occlusion estimation model is introduced to replace the missing pedestrian occlusion annotations in the real-world dataset. Based on a qualitative analysis of the correlation graphs that visualize the correlation between the PLFs and the detection performance, this study identified 16 of the initial 21 factors as being PLFs for DNNs out of which the entropy, the occlusion ratio, the boundary edge strength, and the bounding box aspect ratio turned out to be most severely affecting the detection performance. The findings of this study highlight some of the most serious shortcomings of current DNNs and pave the way for future research to address these issues.

View PDFchevron_right

Related topics

  • Computer Science
  • Artificial Intelligence
  • Computer Vision
  • Pedestrian Detection
    • 580 California St., Suite 400
    San Francisco, CA, 94104
    © 2025 Academia. All rights reserved