dodiku / VNect-tensorflow

VNect -- Tensorflow version

This project is the tensorflow implementation of VNect: Real-time 3D Human Pose Estimation with a Single RGB Camera, SIGGRAPH 2017.

This is not an official implementation. If you find any question regarding to this project please directly contact to me or open an issue.

Environments

• Ubuntu 16.04
• Python 2.7
• Tensorflow 1.3.0
• OpenCV 3.3.0
• OpenGL (optional)

Inference

• 1.Download model, put them in folder models/weights
• 2.Edit demo settings in shell script, --device --demo_type --model_file --test_img --plot_2d --plot_3d
• 3.If you have OpenGL, you can run run_demo_tf_gl.sh for faster rendering of 3d joints. Otherwise, run run_demo_tf.sh

TODO

• Some bugs in detected 3D joint locations.
• Training part of model.