dingyanna / SfSNet-Pytorch

implement SfSNet using pytorch

Introduction

This project partially implement SfSNet project. I will implement all codes in SfSNet project in the future, if I have enough time.

Currently implemented:

• Implement test_SfSNet.m as SfSNet_test.py
• Implement functions/*.m in src/functions.py
• move some constant variables to config.py

If there are bugs in SfSNet_test.py, please open a issue.

Dependencies

• Python libs in requirements.txt

Run SfSNet_test.py

• Download shape_predictor_68_face_landmarks.dat from: http://dlib.net/files/shape_predictor_68_face_landmarks.dat.bz2 , and uncompress it to directory data.

• Create and activate a virtual environment * for python 2.*

  pip install virtualenv
  virtualenv -p python2 venv2
  source venv2/bin/activate

  • for python 3.*

  pip install virtualenv
  virtualenv -p python3 venv3
  source venv3/bin/activate

• Install python dependencies using command:

  pip install -r requirements.txt

• Put your test images in 'Images', and

  python SfSNet_test.py