A problem with video target tracking for key point matching

Question

There is a problem that can't be solved, that is, after obtaining the key points, the tracking based on the key points should be carried out in the subsequent videos. For example, I want to use xfeat to carry out the above tasks, and I also saw the author's video effect display of framing the target and tracking on github, but the author did not release the source code. I'd be grateful if any of the seniors knew how to do that, please! Thanks thanks and thanks!

What I'd like to do is target tracking for videos based on key point matching, as the authors of XFeat: Accelerated Features for Lightweight Image Matching have demonstrated on github