Samurai on segmentation (not-SAM2) video

Hello We love this new tool you developed.

We were wondering if it also works with YOLO on inference / skeleton prediction? We do science research on freely moving rodents and select 6-8 points on animals to track. However, YOLO is very bad at maintaining animal identity over time and swaps between 3 mice in our field-of-view.

Any advice or suggestions are much appreciated!

Thanks so much catubc