Look Around for Anomalies: Weakly-supervised Anomaly Detection via Context-Motion Relational Learning

被引:18
作者
Cho, MyeongAh [1 ]
Kim, Minjung [1 ]
Hwang, Sangwon [2 ]
Park, Chaewon [1 ]
Lee, Kyungjae [3 ]
Lee, Sangyoun [1 ]
机构
[1] Yonsei Univ, Seoul, South Korea
[2] Hyundai Motor Co, Seoul, South Korea
[3] Yong In Univ, Yongin, South Korea
来源
2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) | 2023年
关键词
D O I
10.1109/CVPR52729.2023.01168
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Weakly-supervised Video Anomaly Detection is the task of detecting frame-level anomalies using video-level labeled training data. It is difficult to explore class representative features using minimal supervision of weak labels with a single backbone branch. Furthermore, in real-world scenarios, the boundary between normal and abnormal is ambiguous and varies depending on the situation. For example, even for the same motion of running person, the abnormality varies depending on whether the surroundings are a playground or a roadway. Therefore, our aim is to extract discriminative features by widening the relative gap between classes' features from a single branch. In the proposed Class-Activate Feature Learning (CLAV), the features are extracted as per the weights that are implicitly activated depending on the class, and the gap is then enlarged through relative distance learning. Furthermore, as the relationship between context and motion is important in order to identify the anomalies in complex and diverse scenes, we propose a Context-Motion Interrelation Module (CoMo), which models the relationship between the appearance of the surroundings and motion, rather than utilizing only temporal dependencies or motion information. The proposed method shows SOTA performance on four benchmarks including large-scale real-world datasets, and we demonstrate the importance of relational information by analyzing the qualitative results and generalization ability.
引用
收藏
页码:12137 / 12146
页数:10
相关论文
共 47 条
[1]   Latent Space Autoregression for Novelty Detection [J].
Abati, Davide ;
Porrello, Angelo ;
Calderara, Simone ;
Cucchiara, Rita .
2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019), 2019, :481-490
[2]  
Baradaran, 2022, ARXIV221007697
[3]   TD3 with Reverse KL Regularizer for Offline Reinforcement Learning from Mixed Datasets [J].
Cai, Yuanying ;
Zhang, Chuheng ;
Zhao, Li ;
Shen, Wei ;
Zhang, Xuyun ;
Song, Lei ;
Bian, Jiang ;
Qin, Tao ;
Liu, Tieyan .
2022 IEEE INTERNATIONAL CONFERENCE ON DATA MINING (ICDM), 2022, :21-30
[4]  
Cao Congqi, 2022, ARXIV220206503
[5]   Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset [J].
Carreira, Joao ;
Zisserman, Andrew .
30TH IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2017), 2017, :4724-4733
[6]  
Chalapathy R., 2019, ARXIV190103407, P1
[7]   Anomaly detection in surveillance video based on bidirectional prediction [J].
Chen, Dongyue ;
Wang, Pengtao ;
Yue, Lingyi ;
Zhang, Yuxin ;
Jia, Tong .
IMAGE AND VISION COMPUTING, 2020, 98 (98)
[8]   Graph-Based Global Reasoning Networks [J].
Chen, Yunpeng ;
Rohrbach, Marcus ;
Yan, Zhicheng ;
Yan, Shuicheng ;
Feng, Jiashi ;
Kalantidis, Yannis .
2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019), 2019, :433-442
[9]   Abnormal Event Detection in Videos Using Spatiotemporal Autoencoder [J].
Chong, Yong Shean ;
Tay, Yong Haur .
ADVANCES IN NEURAL NETWORKS, PT II, 2017, 10262 :189-196
[10]   MIST: Multiple Instance Self-Training Framework for Video Anomaly Detection [J].
Feng, Jia-Chang ;
Hong, Fa-Ting ;
Zheng, Wei-Shi .
2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, :14004-14013