3D Face: Fast, Accurate and Stable Reconstruction
This work extends the previous work 3DDFA, named 3DDFA_V2, titled Towards Fast, Accurate and Stable 3D Dense Face Alignment, accepted by ECCV 2020.
3d-face face-aligment computer-vision pytorch notebook paper code arxiv:2009.09960 library research

Introduction

stars

demo

This work extends 3DDFA, named 3DDFA_V2, titled Towards Fast, Accurate and Stable 3D Dense Face Alignment, accepted by ECCV 2020. Compared to 3DDFA, 3DDFA_V2 achieves better performance and stability. Besides, 3DDFA_V2 incorporates the fast face detector FaceBoxes instead of Dlib. A simple 3D render written by c++ and cython is also included. If you are interested in this repo, just try it on this google colab!

Getting started

The usage is very simple.

  1. Clone this repo
git clone https://github.com/cleardusk/3DDFA_V2.git
cd 3DDFA_V2
  1. Build the cython version of NMS, and Sim3DR
sh ./build.sh
  1. Run demos
python3 demo.py -f examples/inputs/emma.jpg  # -o [2d_sparse, 2d_dense, 3d, depth, pncc, pose, uv_tex, ply, obj]
python3 demo_video.py -f examples/inputs/videos/214.avi
python3 demo_video_smooth.py -f examples/inputs/videos/214.avi
python3 demo_webcam_smooth.py

For example, running python3 demo.py -f examples/inputs/emma.jpg -o 3d will give the result below:

demo

More demos:

demo

More features to see here.

Don't forget to tag @cleardusk in your comment, otherwise they may not be notified.

Authors original post
Ph.D. student on face, vision.
Share this project
Similar projects
Face Alignment in Full Pose Range: A 3D Total Solution
Face Alignment in Full Pose Range: A 3D Total Solution.