Blake Blackshear

It could also be that the object isn't a confirmed true positive before it crosses. It takes at least 3 frames where the median score is above the threshold.

I would prefer that each model implementation defines it's own function to prepare the input data in shared memory and then translates the outputs into a common format. Keep in...

I haven't had time to review this in detail yet, but Frigate+ models will be trained on the YUV colorspace, so don't make any assumptions about colorspace here.

This branch is so old that these changes would need to be rewritten on current code.

This looks great. The centroid tracker was always just a first pass. Can you link to any documentation on the SORT tracking algorithm you implemented?

I am in the middle of a few other features at the moment, and I will probably look at incorporating these independently when I am done. Since SORT is GPL-3...

Still planning to revisit the object tracking algorithm.

Finally. I will look into it again.

Not stale

I have seen watsor, and that's similar to the planned (eventually) approach.