Chinese Journal of Liquid Crystals and Displays, Volume. 37, Issue 12, 1607(2022)
Video anomaly detection based on ensemble generative adversarial networks
Fig. 1. Methodology of abnormal event detection based on GAN ensembles
Fig. 2. Anomaly detection comparison on CUHK Avenue Dataset and ShanghaiTech Dataset
Fig. 3. Difference detection performances with different ensemble sizes
Table 1. Comparison of frame-level anomaly detection performance with state-of-the-art methods
View tableView in ArticleTable 1. Comparison of frame-level anomaly detection performance with state-of-the-art methods
Methods CUHK Avenue ShanghaiTech Frame-Pred.［ 17 ］85.1 72.8 VEC［ 18 ］90.2 74.8 Conv-VRNN［ 19 ］85.8 MNAD-P［ 20 ］88.5 70.5 AMDN［ 21 ］84.6 Conv2D-AE［ 6 ］70.2 Conv3D-AE［ 6 ］74.2 StackRNN［ 22 ］80.9 68.0 GAN ensembles 90.6 75.1
Get Citation
Copy Citation Text
Jia-cheng GU, Ying-wen LONG, Ming-ming JI. Video anomaly detection based on ensemble generative adversarial networks[J]. Chinese Journal of Liquid Crystals and Displays, 2022, 37(12): 1607
Category: Research Articles
Received: Apr. 28, 2022
Accepted: --
Published Online: Nov. 30, 2022
The Author Email: Ying-wen LONG (longyingwen@sohu.com)
