Asking for display method

[pogopo789]

First, I would like to extend my sincere gratitude to you for creating this project. The source code you provided is exceptionally clean and accessible, and I truly appreciate your proactive engagement in addressing the issues. Although I encountered several challenges, the guidance provided—though few—was immensely helpful.

Currently, I am working on replicating the results you presented in your paper and on the project's website. I am using Open3D to reconstruct SMPL from video, but the results have not been as clear or effective as yours.

animation.mp4

The NLF can detect 4 people in the video, and I just want to display the first people it detected. But since it detected frame by frame, some frame it got 4 people and some only 3 or maybe it got mismatched in lableing people which leading to the effect you see in the video. Could I have an advice to achieve good result that you have so far. Thank you so much

[pogopo789]