search

HuangJunJie2017 / UDP-Pose

Official code of The Devil is in the Details: Delving into Unbiased Data Processing for Human Pose Estimation
Apache License 2.0
307 stars 54 forks source link

How can i use my own custom dataset and obtain metrics? #16

Closed kathanvyas closed 3 years ago

kathanvyas commented 3 years ago

I have some 500 images and annotation json for those 500 files. I want to evaluate the model on this new test dataset and obtain AP, AP0.5 from that. How can I do that? I could change data dir to my new data directory. However what should I do about the person detection result of COCO val2017 as I do not have that for my dataset for now.

HuangJunJie2017 commented 3 years ago

@kathanvyas you need to prepared data in coco format~ you need to detect human first and save the human bounding boxes in coco format 'person_keypoints_val2017.json' you can refer to mmdet for human detector

kathanvyas commented 3 years ago

Oh so we need to have the human bounding boxes detected prior to using this repo?

HuangJunJie2017 commented 3 years ago

@kathanvyas yes, human bounding boxes is the precondition for top-down paradigm multi-instance human pose estimation but for bottom-down paradigm, human bounding boxes is not necessary. you can refer to HigherHRNet in mmpose for bottom-up methods