How to Create a Standardized Dataset from MP4 Videos After Extracting SMPL?

[superyosan]

Thank you for the excellent code!
I’m planning to use it for my graduation research.
I have a question: Could you explain how to create a dataset from a video (mp4)? Alternatively, if there’s documentation or a guide for this process, could you point me to it?

I’ve used 4D-Humans to extract SMPL, but I’m unsure how to standardize the dataset format. I’ve been looking into repositories like HumanML3D and AMASS, but they seem a bit complex, and I’m currently stuck.

I’d really appreciate your help!