search

mks0601 / 3DMPPE_ROOTNET_RELEASE

Official PyTorch implementation of "Camera Distance-aware Top-down Approach for 3D Multi-person Pose Estimation from a Single RGB Image", ICCV 2019
MIT License
477 stars 65 forks source link

Dataset and visualization? #14

Closed yxxxqqq closed 4 years ago

yxxxqqq commented 4 years ago

  1. how can I get the visualization results? It's seems that there is no visualization code, I want to get the visualization results like your demonstration.

  2. I don't understand the datasets's json files, such as Human36M_subjectx.json， there are some information about camera parameters, but aren't the real coordination (x,y,z) about human? How can I know the real (x, y, z)?

  3. the test result 'bbox_root_human36m_output.json', there are information about 'bbox' and 'root_cam', no information prediction coordination (x, y, z)? How can I know the prediction (x, y, z)?

Sorry to bother you, thank you very much!

mks0601 commented 4 years ago

I think you are interested in https://github.com/mks0601/3DMPPE_POSENET_RELEASE This repo contains RootNet.

yxxxqqq commented 4 years ago

woo, thank you very much !!!