Delayed tracking in the validator?

[Jonathan Aguinaga]

Hello, 

Has anyone noticed any delay in the tracking (observed animal location and expected animal location) when using the validator? I thought it may have been one video, so I tried a different video, but the same issue.

Interestingly, when I watch the video initially, most of it looks great so I scroll through. Then I scroll back to watch it or show a collaborator, but then the tracks are mismatched (Picture below). I wonder if there is something wrong, but I am not exactly sure how to share info about the validator so I attach the log instead.

best,

Jon

[idtrackerai]

Hi Jon,

Sorry for the late reply. The delay you see is something I worked on for a long time, and I still don't know how to fix it entirely...

A technical explanation. OpenCV (the library we use to read the video files) struggles to read some specific videos, I assume because they are somehow corrupted. When this happens, it's really hard to ask OpenCV to read a specific frame in the video.

In the validator we show the video in the background (read from the video file) and the blobs on top (read from the list of blobs in the session folder). The list of blobs is fixed, but the video in the background can be unpredictable since it is being read on the fly by OpenCV and it can misbehave.

I don't know how many users are working with these "corrupted" videos, but I will spend some more time trying to fix this.

Thanks for your feedback,

Jordi

[Jonathan Aguinaga]

Hi Jordi,

No worries! I went on vacation after posting the initial question and just returned a few days ago. So this is very timely for me! 

The technical explanation makes sense to me. I really enjoy the ease of the validator and the video generator in IDtracker. Ive written some code to accomplish the same thing with output from SLEAP and that was a bit of a hassle! So thank you for your continued efforts.

best,

Jon

[idtrackerai]

Thanks!

I started writting the validator for myself when I was an idtrackerai user because I needed to correct the errors in my trackings. I'm happy to share it now as a developer :)

Jordi