shubham-goel/4D-Humans

Evaluation Results for Human 3.6

m-usamasaleem opened this issue · 1 comments

Could you clarify how you collected the images used for testing for human3.6 dataset? The image paths seem to only contain information about the subject and the action. How can I access the same images? Could you also explain if there was a specific pattern or method you used to select these images for testing?

I have all the images for testing, I just want to map your image paths to the original human3.6 testing image path pattern.

Please check the h36m scripts here to see how to map the original Human3.6M data to the format (e.g., naming convention) that we use.