A Survey of object detection in crowded scenes based on deep learning
Haonan Tian
Hunan University of Science and Technology, School of Computer Science and Engineering
DOI: https://doi.org/10.59429/pmcs.v7i2.10323
Keywords: Deep learning; Object detection; Crowded scenes; Non-maximum suppression; YOLO
Abstract
With the progress of deep learning technology, the object detection algorithms have achieved good detection results in general scenes, but they have encountered difficulties in crowded scenes. In crowded scenes, there are a lot of occlusion between objects, which makes the non-maximum suppression algorithm easy to delete the correct detection of overlapping objects; at the same time, there are some problems, such as large change of object scale, small object, insufficient available features and so on. In order to promote the further development of crowded object detection technology, the related methods and techniques are summarized. Firstly, the research background and application value of object detection in crowded scenes are introduced. Secondly, the difficulties of object detection in crowded scenes are discussed, and the defects of object detection algorithm based on deep learning in crowded scenes are analyzed. Then, the existing object detection algorithms in crowded scenes are classified and described. Finally, the possible research direction of object detection in crowded scenes is prospected.
References
[1]Zou Z, Chen K, Shi Z, et al. Object detection in 20 years: A survey[J]. Proceedings of the IEEE, 2023, 111(3): 257-276.
[2]Bodla N, Singh B, Chellappa R, et al. Soft-NMS--improving object detection with one line of code[C]//Proceedings of the IEEE international conference on computer vision. 2017: 5561-5569.
[3]Liu S, Huang D, Wang Y. Adaptive nms: Refining pedestrian detection in a crowd[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2019: 6459-6468.
[4]Wang X, Xiao T, Jiang Y, et al. Repulsion loss: Detecting pedestrians in a crowd[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 7774-7783.
[5]Zhang S, Wen L, Bian X, et al. Occlusion-aware R-CNN: detecting pedestrians in a crowd[C]//Proceedings of the European conference on computer vision (ECCV). 2018: 637-653.
[6]Gao Q, Tang F, Li D, et al. Research on pedestrian detection method in dense scene based on improved YOLOv5. Foreign Electronic Measurement Technology. 2023;42(125-130).
[7]Liu Z, Xu H, Zhu X, et al. Bi-YOLO: An Improved Lightweight Object Detection Algorithm Based on YOLOv8. Computer Engineering Science.(1-15).