AnomalyDetection_CVPR18
AnomalyDetection_CVPR18 copied to clipboard
Anomaly detection in real time
How can we make it detect anomaly in real time? will we still need extrapolation of the video frames? should i reduce no of frames per segment to 1?
I would also like to know this? especially if anyone had got it working before and can tell me the kind of performance I can expect from it in real-time? (https://github.com/ekosman/AnomalyDetectionCVPR2018-Pytorch/issues/16 also asked here)