The project is an official implementation of our CVPR2019 paper "Deep High-Resolution Representation Learning for Human Pose Estimation"
4.31k
stars
908
forks
source link
代码bug,USE_GT_BBOX 参数设置为false时,使用detector的bbox,没有读取keypoints #276
Open
SherlockUnknowEn opened 2 years ago
测试时,当yaml 中USE_GT_BBOX为false时,![捕获](https://user-images.githubusercontent.com/13676092/156799621-54b8dc7b-df23-433b-ad72-11ab7da4e95f.PNG)
lib/dataset/coco.py
中 136行gt_db = self._load_coco_person_detection_results()
,_load_coco_person_detection_results 函数里只读取了 bbox,没有读取关键点,请问是否是代码bug?