ABSTRACT
In urban surveillance systems, finding a specific vehicle in video frames efficiently and accurately has always been an essential part of traffic supervision and criminal investigation. Existing studies focus on vehicle re-identification (re-ID), but vehicle search is still underexploited. These methods depend on the locations of many vehicles (bounding boxes) that are not available in most real-world applications. Therefore, the unsupervised joint study of vehicle location and identification for the observed scene is a pressing need. Inspired by person search, we conduct a study on the vehicle search while considering four main discrepancies among them, summarized as: 1) It is challenging to select the candidate regions for the observed vehicle due to the perspective differences (front or side); 2) The sides of the same type of vehicles are almost the same, resulting in smaller inter-class; 3) Lacking satisfied dataset for vehicle search to meet the practical scenarios; 4) Supervised search publishing methods rely on datasets with expensive annotations. To address these issues, we have established a new vehicle search dataset. We design an unsupervised framework on this benchmark dataset to generate pseudo labels for further training existing vehicle re-ID or person search models. Experimental results reveal that these methods turn less effective on vehicle search tasks. Therefore, the vehicle search task needs to be further developed, and this dataset can advance the research of vehicle search. Https://github.com/zsl1997/VSW.
Supplemental Material
- Yan Bai, Yihang Lou, Yongxing Dai, Jun Liu, Ziqian Chen, and Ling-Yu Duan. 2020. Disentangled Feature Learning Network for Vehicle Re-Identification. In Proc. Int. Joint Conf. Artif. Intell. (IJCAI). 474--480.Google ScholarCross Ref
- Raja Muhammad Saad Bashir, Muhammad Shahzad, and M. M. Fraz. 2019. VR-PROUD: Vehicle Re-identification using PROgressive Unsupervised Deep architecture. Pattern Recognit., Vol. 90 (2019), 52--65.Google ScholarCross Ref
- Alexey Bochkovskiy, Chien-Yao Wang, and Hong-Yuan Mark Liao. 2020. YOLOv4: Optimal Speed and Accuracy of Object Detection. arXiv:2004.10934 (2020). https://arxiv.org/abs/2004.10934Google Scholar
- Di Chen, Shanshan Zhang, Wanli Ouyang, Jian Yang, and Bernt Schiele. 2020 a. Hierarchical Online Instance Matching for Person Search. In Proc. AAAI Conf. Artif. Intell. 10518--10525.Google ScholarCross Ref
- Di Chen, Shanshan Zhang, Wanli Ouyang, Jian Yang, and Ying Tai. 2018. Person Search via a Mask-Guided Two-Stream CNN Model. In Proc. Springer Eur. Conf. Comput. Vis. (ECCV). 764--781.Google ScholarCross Ref
- Di Chen, Shanshan Zhang, Jian Yang, and Bernt Schiele. 2020 b. Norm-Aware Embedding for Efficient Person Search. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 12612--12621.Google ScholarCross Ref
- Ruihang Chu, Yifan Sun, Yadong Li, Zheng Liu, Chi Zhang, and Yichen Wei. 2019. Vehicle Re-Identification With Viewpoint-Aware Metric Learning. In Proc. IEEE/CVF Int. Conf. Comput. Vis. (ICCV). 8281--8290.Google ScholarCross Ref
- Jiansheng Dong, Jingling Yuan, Lin Li, and Xian Zhong. 2020 a. A Lightweight High-Resolution Representation Backbone For Real-Time Keypoint-Based Object Detection. In Proc. IEEE Int. Conf. Multimedia Expo (ICME). 1--6.Google ScholarCross Ref
- Jiansheng Dong, Jingling Yuan, Lin Li, Xian Zhong, and Weiru Liu. 2020 b. Optimizing Queries over Video via Lightweight Keypoint-based Object Detection. In Proc. ACM Int. Conf. Multimedia Retr. (ICMR). 548--554. Google ScholarDigital Library
- Wenkai Dong, Zhaoxiang Zhang, Chunfeng Song, and Tieniu Tan. 2020 c. Bi-Directional Interaction Network for Person Search. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 2836--2845.Google ScholarCross Ref
- Wenkai Dong, Zhaoxiang Zhang, Chunfeng Song, and Tieniu Tan. 2020 d. Instance Guided Proposal Network for Person Search. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 2582--2591.Google ScholarCross Ref
- Bing He, Jia Li, Yifan Zhao, and Yonghong Tian. 2019. Part-Regularized Near-Duplicate Vehicle Re-Identification. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 3997--4005.Google ScholarCross Ref
- Kaiming He, Georgia Gkioxari, Piotr Dollá r, and Ross B. Girshick. 2017. Mask R-CNN. In Proc. IEEE/CVF Int. Conf. Comput. Vis. (ICCV). 2980--2988.Google Scholar
- Shuting He, Hao Luo, Weihua Chen, Miao Zhang, Yuqi Zhang, Fan Wang, Hao Li, and Wei Jiang. 2020. Multi-Domain Learning and Identity Mining for Vehicle Re-Identification. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR) Workshops. 2485--2493.Google ScholarCross Ref
- Ziling Huang, Zheng Wang, Chung-Chi Tsai, Shin'ichi Satoh, and Chia-Wen Lin. 2021. DotSCN: Group Re-Identification via Domain-Transferred Single and Couple Representation Learning. IEEE Trans. Circuits Syst. Video Technol., Vol. 31, 7 (2021), 2739--2750.Google ScholarCross Ref
- Wei Li, Shaogang Gong, and Xiatian Zhu. 2021. Hierarchical distillation learning for scalable person search. Pattern Recognit., Vol. 114 (2021), 107862.Google ScholarCross Ref
- Wei Li, Zhenting Wang, Xiao Wu, Ji Zhang, Qiang Peng, and Hongliang Li. 2020. CODAN: Counting-driven Attention Network for Vehicle Detection in Congested Scenes. In Proc. ACM Int. Conf. Multimedia (ACM MM). 73--82. Google ScholarDigital Library
- Tsung-Yi Lin, Michael Maire, Serge J. Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Dollá r, and C. Lawrence Zitnick. 2014. Microsoft COCO: Common Objects in Context. In Proc. Springer Eur. Conf. Comput. Vis. (ECCV). 740--755.Google Scholar
- Hongye Liu, Yonghong Tian, Yaowei Wang, Lu Pang, and Tiejun Huang. 2016b. Deep Relative Distance Learning: Tell the Difference between Similar Vehicles. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 2167--2175.Google ScholarCross Ref
- Jiawei Liu, Zheng-Jun Zha, Richang Hong, Meng Wang, and Yongdong Zhang. 2020 b. Dual Context-Aware Refinement Network for Person Search. In Proc. ACM Int. Conf. Multimedia (ACM MM). 3450--3459. Google ScholarDigital Library
- Wu Liu, Xinchen Liu, Huadong Ma, and Peng Cheng. 2017a. Beyond Human-level License Plate Super-resolution with Progressive Vehicle Search and Domain Priori GAN. In Proc. ACM Int. Conf. Multimedia (ACM MM). 1618--1626. Google ScholarDigital Library
- Wu Liu, Xinchen Liu, Huadong Ma, and Peng Cheng. 2017b. Beyond Human-level License Plate Super-resolution with Progressive Vehicle Search and Domain Priori GAN. In Proc. ACM Int. Conf. Multimedia (ACM MM). 1618--1626. Google ScholarDigital Library
- Xinchen Liu, Wu Liu, Huadong Ma, and Huiyuan Fu. 2016a. Large-scale vehicle re-identification in urban surveillance videos. In Proc. IEEE Int. Conf. Multimedia Expo (ICME). 1--6.Google ScholarCross Ref
- Xinchen Liu, Wu Liu, Jinkai Zheng, Chenggang Yan, and Tao Mei. 2020 a. Beyond the Parts: Learning Multi-view Cross-part Correlation for Vehicle Re-identification. In Proc. ACM Int. Conf. Multimedia (ACM MM). 907--915. Google ScholarDigital Library
- Xinchen Liu, Huadong Ma, and Shuangqun Li. 2019. PVSS: A Progressive Vehicle Search System for Video Surveillance Networks. J. Comput. Sci. Technol., Vol. 34, 3 (2019), 634--644.Google ScholarCross Ref
- Yihang Lou, Yan Bai, Jun Liu, Shiqi Wang, and Lingyu Duan. 2019. VERI-Wild: A Large Dataset and a New Method for Vehicle Re-Identification in the Wild. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 3235--3243.Google ScholarCross Ref
- Dechao Meng, Liang Li, Xuejing Liu, Yadong Li, Shijie Yang, Zheng-Jun Zha, Xingyu Gao, Shuhui Wang, and Qingming Huang. 2020. Parsing-Based View-Aware Embedding Network for Vehicle Re-Identification. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 7101--7110.Google ScholarCross Ref
- Curtis G. Northcutt, Anish Athalye, and Jonas Mueller. 2021. Pervasive Label Errors in Test Sets Destabilize Machine Learning Benchmarks. arXiv:2103.14749 (2021). https://arxiv.org/abs/2103.14749Google Scholar
- Jinjia Peng, Yang Wang, Huibing Wang, Zhao Zhang, Xianping Fu, and Meng Wang. 2020. Unsupervised Vehicle Re-identification with Progressive Adaptation. In Proc. Int. Joint Conf. Artif. Intell. (IJCAI). 913--919.Google ScholarCross Ref
- Ziruo Sun, Xiushan Nie, Xiaoming Xi, and Yilong Yin. 2020. CFVMNet: A Multi-branch Network for Vehicle Re-identification Based on Common Field of View. In Proc. ACM Int. Conf. Multimedia (ACM MM). 3523--3531. Google ScholarDigital Library
- Mingxing Tan, Ruoming Pang, and Quoc V. Le. 2020. EfficientDet: Scalable and Efficient Object Detection. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 10778--10787.Google Scholar
- Zheng Tang, Milind Naphade, Ming-Yu Liu, Xiaodong Yang, Stan Birchfield, Shuo Wang, Ratnesh Kumar, David C. Anastasiu, and Jenq-Neng Hwang. 2019. CityFlow: A City-Scale Benchmark for Multi-Target Multi-Camera Vehicle Tracking and Re-Identification. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 8797--8806.Google ScholarCross Ref
- Cheng Wang, Bingpeng Ma, Hong Chang, Shiguang Shan, and Xilin Chen. 2020 d. TCTS: A Task-Consistent Two-Stage Framework for Person Search. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 11949--11958.Google ScholarCross Ref
- Xiao Wang, Jun Chen, Zheng Wang, Wu Liu, Shin'ichi Satoh, Chao Liang, and Chia-Wen Lin. 2020 a. When Pedestrian Detection Meets Nighttime Surveillance: A New Benchmark. In Proc. Int. Joint Conf. Artif. Intell. (IJCAI). 509--515.Google ScholarCross Ref
- Xiao Wang, Chao Liang, Chen Chen, Jun Chen, Zheng Wang, Zhen Han, and Chunxia Xiao. 2020 b. S3D: Scalable Pedestrian Detection via Score Scale Surface Discrimination. IEEE Trans. Circuits Syst. Video Technol., Vol. 30, 10 (2020), 3332--3344.Google ScholarCross Ref
- Zheng Wang, Wu Liu, Yusuke Matsui, and Shin'ichi Satoh. 2020 c. Effective and Efficient: Toward Open-world Instance Re-identification. In Proc. ACM Int. Conf. Multimedia (ACM MM). 4789--4790. Google ScholarDigital Library
- Zheng Wang, Zhixiang Wang, Yinqiang Zheng, Yang Wu, Wenjun Zeng, and Shin'ichi Satoh. 2020 e. Beyond Intra-modality: A Survey of Heterogeneous Person Re-identification. In Proc. Int. Joint Conf. Artif. Intell. (IJCAI). 4973--4980.Google ScholarCross Ref
- Nicolai Wojke, Alex Bewley, and Dietrich Paulus. 2017. Simple online and realtime tracking with a deep association metric. In Proc. IEEE Int. Conf. Image Process. (ICIP). 3645--3649.Google ScholarDigital Library
- Tong Xiao, Shuang Li, Bochao Wang, Liang Lin, and Xiaogang Wang. 2017. Joint Detection and Identification Feature Learning for Person Search. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 3376--3385.Google ScholarCross Ref
- Pengyu Xie, Xin Xu, Zheng Wang, and Toshihiko Yamasaki. 2021. Unsupervised Video Person Re-identification via Noise and Hard frame Aware Clustering. arXiv:2106.05441 (2021). https://arxiv.org/abs/2106.05441Google Scholar
- Xin Xu, Lei Liu, Xiaolong Zhang, Weili Guan, and Ruimin Hu. 2021. Rethinking data collection for person re-identification: active redundancy reduction. Pattern Recognit., Vol. 113 (2021), 107827.Google ScholarCross Ref
- Ke Yan, Yonghong Tian, Yaowei Wang, Wei Zeng, and Tiejun Huang. 2017. Exploiting Multi-grain Ranking Constraints for Precisely Searching Visually-similar Vehicles. In Proc. IEEE/CVF Int. Conf. Comput. Vis. (ICCV). 562--570.Google ScholarCross Ref
- Yichao Yan, Jingpeng Li, Jie Qin, Song Bai, Shengcai Liao, Li Liu, Fan Zhu, and Ling Shao. 2021. Anchor-Free Person Search. arXiv:2103.11617 (2021). https://arxiv.org/abs/2103.11617Google Scholar
- Wenjie Yang, Dangwei Li, Xiaotang Chen, and Kaiqi Huang. 2020. Bottom-Up Foreground-Aware Feature Fusion for Person Search. In Proc. ACM Int. Conf. Multimedia (ACM MM). 3404--3412. Google ScholarDigital Library
- Hantao Yao and Changsheng Xu. 2021. Joint Person Objectness and Repulsion for Person Search. IEEE Trans. Image Process., Vol. 30 (2021), 685--696.Google ScholarDigital Library
- Yue Yao, Liang Zheng, Xiaodong Yang, Milind Naphade, and Tom Gedeon. 2020. Simulating Content Consistent Vehicle Datasets with Attribute Descent. In Proc. Springer Eur. Conf. Comput. Vis. (ECCV). 775--791.Google ScholarCross Ref
- Jongmin Yu and Hyeontaek Oh. 2021. Unsupervised Vehicle Re-Identification via Self-supervised Metric Learning using Feature Dictionary. arXiv:2103.02250 (2021). https://arxiv.org/abs/2103.02250Google Scholar
- Lei Zhang, Zhenwei He, Yi Yang, Liang Wang, and Xinbo Gao. 2020. Tasks Integrated Networks: Joint Detection and Retrieval for Image Search. arXiv:2009.01438 (2020). https://arxiv.org/abs/2009.01438Google Scholar
- Liang Zheng, Hengheng Zhang, Shaoyan Sun, Manmohan Chandraker, Yi Yang, and Qi Tian. 2017. Person Re-identification in the Wild. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR). 3346--3355.Google ScholarCross Ref
- Xian Zhong, Meng Feng, Wenxin Huang, Zheng Wang, and Shin'ichi Satoh. 2019. Poses Guide Spatiotemporal Model for Vehicle Re-identification. In Proc. Springer Int. Conf. Multim. Model. (MMM). 426--439.Google ScholarCross Ref
- Xian Zhong, Yiting Liu, Wenxin Huang, Xiao Wang, Bo Ma, and Jingling Yuan. 2021 a. Part-aligned Network with Background for Misaligned Person Search. Proc. IEEE Int. Conf. Acoustics Speech Signal Process. (ICASSP), 4250--4254.Google ScholarCross Ref
- Xian Zhong, Tianyou Lu, Wenxin Huang, Mang Ye, and Chia-Wen Lin. 2021 b. Grayscale Enhancement Colorization Network for Visible-infrared Person Re-identification. IEEE Trans. Circuits Syst. Video Technol. (2021).Google Scholar
- Wenqian Zhu, Ruimin Hu, Zhongyuan Wang, Dengshi Li, and Xiyue Gao. 2020 a. Tell The Truth From The Front: Anti-Disguise Vehicle Re-Identification. In Proc. IEEE Int. Conf. Multimedia Expo (ICME). 1--6.Google ScholarCross Ref
- Yangchun Zhu, Zheng-Jun Zha, Tianzhu Zhang, Jiawei Liu, and Jiebo Luo. 2020 b. A Structured Graph Attention Network for Vehicle Re-Identification. In Proc. ACM Int. Conf. Multimedia (ACM MM). 646--654. Google ScholarDigital Library
Index Terms
- Unsupervised Vehicle Search in the Wild: A New Benchmark
Recommendations
Vehicle Re-Identification Based on Unsupervised Domain Adaptation by Incremental Generation of Pseudo-Labels
Progress in Pattern Recognition, Image Analysis, Computer Vision, and ApplicationsAbstractThe main goal of vehicle re-identification (ReID) is to associate the same vehicle identity in different cameras. This is a challenging task due to variations in light, viewpoints or occlusions; in particular, vehicles present a large intra-class ...
DUSA: Decoupled Unsupervised Sim2Real Adaptation for Vehicle-to-Everything Collaborative Perception
MM '23: Proceedings of the 31st ACM International Conference on MultimediaVehicle-to-Everything (V2X) collaborative perception is crucial for the advancement of autonomous driving. However, achieving high-precision V2X perception requires a significant amount of annotated real-world data, which can always be expensive and hard ...
Unsupervised Vehicle Re-Identification via Raw UAV Videos
Image and GraphicsAbstractFor matching vehicles across different camera views, vehicle Re-Identification has made great progress in supervised learning. However, supervised approach would require extensive manual labeling which is costly and unfeasible for large-scale ...
Comments