OCSort non-default max_age error

[MABatin]

Search before asking

• I have searched the Yolo Tracking issues and discussions and found no similar questions.

Yolo Tracking Component

Tracking

Bug

When increasing the value of max_age parameter of OCSort class, I get the following error:

Traceback (most recent call last):
  File "/home/ml-machine/batin/easy-security_onprem/unused/tracker-tests/ocsort.py", line 66, in <module>
    main(parse_args())
  File "/home/ml-machine/batin/easy-security_onprem/unused/tracker-tests/ocsort.py", line 52, in main
    tracks = tracker.update(dets, im) # --> M X (x, y, x, y, id, conf, cls, ind)
  File "/home/ml-machine/miniconda3/envs/ublenv/lib/python3.10/site-packages/boxmot/utils/__init__.py", line 79, in wrapper
    tracks = self.update(instance, dets, im)
  File "/home/ml-machine/miniconda3/envs/ublenv/lib/python3.10/site-packages/boxmot/trackers/ocsort/ocsort.py", line 292, in update
    self.active_tracks[m[1]].update(dets[m[0], :5], dets[m[0], 5], dets[m[0], 6])
  File "/home/ml-machine/miniconda3/envs/ublenv/lib/python3.10/site-packages/boxmot/trackers/ocsort/ocsort.py", line 167, in update
    self.kf.update(convert_bbox_to_z(bbox))
  File "/home/ml-machine/miniconda3/envs/ublenv/lib/python3.10/site-packages/boxmot/motion/kalman_filters/ocsort_kf.py", line 485, in update
    self.unfreeze()
  File "/home/ml-machine/miniconda3/envs/ublenv/lib/python3.10/site-packages/boxmot/motion/kalman_filters/ocsort_kf.py", line 399, in unfreeze
    index1 = indices[-2]
IndexError: index -2 is out of bounds for axis 0 with size 1

Environment

Boxmot - 10.0.65
Python - 3.10.13 torch 2.1.2+cu118 CUDA:0 (RTX 3090 24GB)

Minimal Reproducible Example

    detector = YOLO(
        model="path/to/yolov8m.pt"
    ).to(device)
    tracker = OCSORT(
        max_age=60
    )
    kwargs = {'classes': 0}
    results = detector(input_filename, stream=True, **kwargs)

    for r in results:
        try:
            im = deepcopy(r.orig_img)
            annotator = Annotator(im)
            dets = r.boxes.data.cpu().numpy()
            tracks = tracker.update(dets, im) # --> M X (x, y, x, y, id, conf, cls, ind)
            for track in tracks:
                ...

[mikel-brostrom]

So it works with 30?

[J4BEZ]

I got same error while using HybridSORT.

in my case, it works very well with 30 and 40.
but if I increase max_age to 120, I got the same error occured😥

[mikel-brostrom]

I see. The max deque length for observation history is 50 so it makes sense that it works with 30 and 40, but not over 50

[MABatin]

I see. The max deque length for observation history is 50 so it makes sense that it works with 30 and 40, but not over 50

Are you planning on making the max deque length dynamic as the max_age or could there be any downside to it?

[github-actions]

👋 Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.
Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

[mikel-brostrom]

I see. The max deque length for observation history is 50 so it makes sense that it works with 30 and 40, but not over 50

Are you planning on making the max deque length dynamic as the max_age or could there be any downside to it?

Yup, I have it in my backlog 😄

[github-actions]

👋 Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.
Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!