search

henrryzh1 / UR-DMU

Official code for AAAI2023 paper "Dual Memory Units with Uncertainty Regulation for Weakly Supervised Video Anomaly Detection"
MIT License
65 stars 5 forks source link

about the "gt" generate #6

Closed YukiFan closed 1 year ago

YukiFan commented 1 year ago

sorry to bother you again...... in readme.md about feature_extract, you say image but I wonder if the feature size is 101, but the gt size is 100, isn't inconsistency during testing? thank you~~

henrryzh1 commented 1 year ago

For this part, I did not describe very well. There are two cases to get the features, drop the last frames or keep last frames. We use the first case for GT and feature extraction. 1601-->1600-->100 clips. But the code of feature extraction we provide is the keeping last frames case. If you use the feature extraction code, the GT needs to be regenerated. You can also change the code to be the dropping last frames case.

YukiFan commented 1 year ago

OK!!!GET!!!THANKYOU!!！