search

fjchange / object_centric_VAD

An Tensorflow Re-Implement of CVPR 2019 "Object-centric Auto-Encoders and Dummy Anomalies for Abnormal Event Detection in Video"
MIT License
95 stars 30 forks source link

AUC computation #13

Closed kevaldoshi17 closed 4 years ago

kevaldoshi17 commented 4 years ago

Hello,

You say that the VAD authors compute the AUC by first computing it per video and then averaging it but I could not find it mentioned anywhere. Could you please point me where it is mentioned?

Thank You