Can't reproduce the result on Avenue dataset reported in the paper.
Closed this issue · 4 comments
Thank you for your open source of your great work!
When I run the expriment on the Avenue dataset,I have the following problem:
After I running the Train_Avenue.py file,I get the best auc score is 81.54,which is far lower than the result reported in your paper.
Can you tell me how to reproduce the auc score of 92.8 in your paper.
Hope for your reply,tanks!
I can reproduce the results of the ped2 and shanghai datasets, but even if the pretrained model you provided is loaded on the avenue dataset, only 81% of the AUC results are obtained. May I ask what is the reason for this? I hope to receive your reply. Thank you!
avenue dataset needs extra post-processing, as DMAD could identify static anomalies where labeled as normal in original annotation (e.g. a bag in the bottom left corner of the first few frames)
Please run process_avenue.py to remove these static anomalies.
avenue dataset needs extra post-processing, as DMAD could identify static anomalies where labeled as normal in original annotation (e.g. a bag in the bottom left corner of the first few frames)
Please run process_avenue.py to remove these static anomalies.
I have solved the problem.Thank you!
avenue dataset needs extra post-processing, as DMAD could identify static anomalies where labeled as normal in original annotation (e.g. a bag in the bottom left corner of the first few frames)
Please run process_avenue.py to remove these static anomalies.I have solved the problem.Thank you!
: )