search

cheind / py-motmetrics

:bar_chart: Benchmark multiple object trackers (MOT) in Python
MIT License
1.37k stars 259 forks source link

how can I get the MOT metrics when using custom dataset(COCO or VOC format) as training set? #119

Closed Veronica1997 closed 4 years ago

Veronica1997 commented 4 years ago

Should I convert the COCO or VOC format datasets to MOT format? Do you know how to convert? Thanks a lot!

cheind commented 4 years ago

Hey, I believe I have seen this done before. Facebook research provides this code I guess:

https://github.com/facebookresearch/DetectAndTrack/

[...] Since the datasets are represented using COCO conventions in Detectron code base, we need the COCO API to be able to read the train/test files. [...] Evaluation is performed using the official PoseTrack evaluation code, poseval, which uses py-motmetrics internally

Veronica1997 commented 4 years ago

Hey, I believe I have seen this done before. Facebook research provides this code I guess:

https://github.com/facebookresearch/DetectAndTrack/

[...] Since the datasets are represented using COCO conventions in Detectron code base, we need the COCO API to be able to read the train/test files. [...] Evaluation is performed using the official PoseTrack evaluation code, poseval, which uses py-motmetrics internally

Thank you very much! It completely solved my problem! appreciate!

cheind commented 4 years ago

Great!