This repository is an official PyTorch implementation of the paper "Learnable Triangulation of Human Pose" (ICCV 2019, oral). Proposed method archives state-of-the-art results in multi-view 3D human pose estimation!
Hello, thank's for sharing the code of this project.
I have some questions :
1./ Possibility estimate 3D pose for a custom image (internet image) (should we have at least a 4 views to the person an each image ?)
2./ How it's the number of FPS for the hardware used .
3./ An the paper evaluation (MPJPE =17.7) use 2D keypoint Ground thruth ? Any idea how the value using only the 2D keypoint detector.
Hello, thank's for sharing the code of this project.
I have some questions :
1./ Possibility estimate 3D pose for a custom image (internet image) (should we have at least a 4 views to the person an each image ?) 2./ How it's the number of FPS for the hardware used . 3./ An the paper evaluation (MPJPE =17.7) use 2D keypoint Ground thruth ? Any idea how the value using only the 2D keypoint detector.
Thank's for your reply .