MOT postprocessing of drawing frames of trackings and thereby a video

[shuyueW1991]

I really appreciate the work of X-AnyLabeling, which is MARVELLOUS. However, i find the postprocessing of MOT detection and tracking video should be added. Therefore I write a small snippet so that after MOT operation, you can run this script and you can get a movie.mp4 registering trackings of objects.

The code is here:

A capture of such image is like this:

There are two editable parts in the script: The first input is the folder path of the frames and the csv filer path (which is default under folder ). These can be edited under the line if __name == '__main__': The second is at the beginning of code, i.e. index2color dict, where you should assign at least the index-color pairs as many as the most id one frame can get. And there are still some comments that you can reference.

Anyway, I hope everyone encountering same problem can use it. Cheers.

[shuyueW1991]

添加了多目标跟踪视频的后处理 , 写了一个小程序使您可以在 MOT 操作之后运行此脚本，并获得记录对象轨迹的 movie.mp4

[CVHub520]

添加了多目标跟踪视频的后处理 , 写了一个小程序使您可以在 MOT 操作之后运行此脚本，并获得记录对象轨迹的 movie.mp4

Thank you for your valuable contribution to X-AnyLabeling!