sangyun884 / HR-VITON

Official PyTorch implementation for the paper High-Resolution Virtual Try-On with Misalignment and Occlusion-Handled Conditions (ECCV 2022).
831 stars 171 forks source link

Openpose body25 #26

Closed tiamo405 closed 1 year ago

tiamo405 commented 1 year ago

When making the code, I want to use my poses that I created myself, but online it's mostly 18 point bodies. I want the output like the article, what code should I use. thank you

martacosano commented 1 year ago

Hello @tiamo405 ! you can use OpenPose with configuration BODY_25 (25 keypoints) to extract the pose, instead of coco (18 keypoints).Here there's a photo of both configurations: https://www.researchgate.net/figure/Architecture-of-the-different-modules-to-recognise-poses-and-control-the-robot_fig3_340618434

tiamo405 commented 1 year ago

Hello @tiamo405 ! you can use OpenPose with configuration BODY_25 (25 keypoints) to extract the pose, instead of coco (18 keypoints).Here there's a photo of both configurations: https://www.researchgate.net/figure/Architecture-of-the-different-modules-to-recognise-poses-and-control-the-robot_fig3_340618434

thank you. can you give me a reference link to body_25, at github openpos I can't install it to use. is there any more convenient way, i have also used points of body_18 to form body_25 but when it becomes input the output fails. I don't know if my current error is in the pose or the points.

tuyendb commented 1 year ago

could u give me github url ?

koo616 commented 1 year ago

Hello guys. You can check the detailed command line operation I used for the openpose. Please check it :)