Closed maoyunyao closed 2 years ago
Basically, I extracted the individual actions using temporal detection annotations and then treat them as independent videos. Then, you can follow the same procedure as given for NTU. I will have a look and see if can find the PKU MMD processing.
It works! Thanks a lot~
Sorry for troubling again! I tried to download Phase #1 of the PKU-MMD dataset at https://www.icst.pku.edu.cn/struct/Projects/PKUMMD.html but the google-drive seems doesn't work for now. If possible, could you kindly share the Phase #1 data with me? Many thanks.
Hello, could you share the processed videos of PKU-MMD? Thanks a lot.
Basically, I extracted the individual actions using temporal detection annotations and then treat them as independent videos. Then, you can follow the same procedure as given for NTU. I will have a look and see if can find the PKU MMD processing.
@maoyunyao you can download pkummd datasets from https://www.icst.pku.edu.cn/struct/Projects/PKUMMD.html I just confirmed it. @dingli93 I don't have any space to do it yet. Will see if I can find any.
Nice work! I am currently following your skeleton-contrast and trying to run it on the PKU MMD dataset. May I have a copy of the data processing code and the training/testing configurations? Thanks in advance~