Ice

how do you call the script?

No idea. Try to follow the guide written in the repo: 1. create database folder using the file structures i listed. 2. make sure you pass the correct parameters to...

just ignore that. thanks :)

Can you specify what's > optical strain codes about?

You can compute using the `output` from `output = readFlowFile(target_flow)` in External-Tools/tvl1flow/output_flo.m Just compute exx, eyy, exy and eyx as written in the paper. I forgot to upload the codes....

1. need to send via email. 2. yea. you can create account from megvii face++ and copy pasta the key. 3. yea. you have to tinker with the code though...

Please refer https://github.com/IcedDoggie/Micro-Expression-with-Deep-Learning/issues/26

I think reshape is fine. I remember the purpose of this is because my keras accepts "channel first". Reshape should be ok all the while. Any issues on your side?

Hi, i added the argument to train_samm_cross.py Thanks for raising out. Please verify.

Hi there, What are "real scenes" ? Currently the best method is around 0.7 in terms of F1 score and it seems that more people are participating in this field....