I noticed that all frames of the 8th and 18th test videos in Avenue dataset are all anomalies. Could you explain how to calculate the AUC of these two videos?
We noticed that and to overcame this issue, we pad the videos with one "0" value at the beginning and one "1" value at the end. In this way you can compute the AUC without any error.
I noticed that all frames of the 8th and 18th test videos in Avenue dataset are all anomalies. Could you explain how to calculate the AUC of these two videos?