MTFL: Multi-Timescale Feature Learning for Weakly-Supervised Anomaly Detection in Surveillance Videos

Benchmark Model Rank Results
anomaly-detection-in-surveillance-videos-on-ucf-crimeMTFL (VST, finetuned on VADD)#2ROC AUC: 89.78
anomaly-detection-in-surveillance-videos-on-ucf-crimeMTFL (VST)#5ROC AUC: 87.16
anomaly-detection-in-surveillance-videos-on-xd-violenceMTFL (VST)#5AP: 84.57
anomaly-detection-in-surveillance-videos-on-xd-violenceMTFL (VST, finetuned on VADD)#11AP: 79.40
anomaly-detection-in-surveillance-videos-on-shanghaitech-weakly-supervisedMTFL (VST, finetuned on VADD)#7AUC-ROC: 95.70
anomaly-detection-in-surveillance-videos-on-shanghaitech-weakly-supervisedMTFL (VST)#8AUC-ROC: 95.32