keelia-python-eos

An implementation of modified-eos using python

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.

Prerequisites

What things you need to install the software and how to install them

dlib
cv2
numpy
image-cleaner
modified-eos

How to setup modified-eos

$ clone https://github.com/xiaoyuetang/eos.git
$ cd eos && python setup.py install

Usage

PLEASE CHANGE FILE PATH IN THE .PY FILES BEFORE YOU PROCEED

$ python generate_faces.py to generate cropped faces with size 240

$ image-cleaner faces_result to remove duplicated images in faces_result

$ python generate_pts.py to generate landmarks for each cropped face

$ python ground_truth.py to generate .npy file storing dictionary of groud truth values:

{"tx": tx, "ty": ty, "scale": scale, "yaw": yaw, "roll": roll, "pitch": pitch, "pca_shape_coefficients": pca_shape_coefficients, "expression_coefficients": expression_coefficients}

$ python train.py to train fine-tuned resnet18 to generate ground truth

Build With

eos - A lightweight 3D Morphable Face Model fitting library in modern C++14
Pytorch - An open source machine learning library based on the Torch library

xiaoyuetang / keelia-python-eos