Makehuman-face-generator-plugin

This plugin utilises the python face recognition library, to automatically create faces from reference images. The project is based upon the face_recognition library (https://github.com/ageitgey/face_recognition), as well as makehuman.

Installation

Requirements

• Python 3.5+
• pip

Installation Options:

Installing on Mac or Linux

A virtual enviroment (venv,anaconda) is recommended but not necessary)

• See here for Install Instructions: venv, anaconda
• Install dlib (GPU support is highly recommended)

Install the requirements using pip

	pip install -r requirements.txt

Clone the source code version of makehuman from here

• Move into the cloned repository.
• Add the 0_modeling_facerec.py from this repository to the makehuman/plugins folder.
• Run the makehuman project (in your virtual enviroment) using:

	python makehuman/makehuman.py

Installing on Windows:

• A virtual enviroment (anaconda) is recommended but not necessary)
• See here for Install Instructions: anaconda

Install dlib (GPU support is highly recommended)

Installing Dlib on windows is a bit more complicated than on linux or mac.

• Here is an install instruction
• Another instruction from the face recognition page

Install the requirements using pip

	pip install -r requirements.txt

Clone the source code version of makehuman from here

• Move into the cloned repository.
• Add the 0_modeling_facerec.py from this repository to the makehuman/plugins folder.
• Run the makehuman project (in your virtual enviroment) using:

	python makehuman/makehuman.py

Usage:

Currently the preview is rendered in a seperate windows which can be hidden behind the main window.

• After starting makehuman a new menu should have appeared:

The "Face Reference Directory" opens a folder dialog window, which should be pointed to the reference images.

• Attention: The Images are loaded from all subfolders.

Multiple different optimizer methods can be chosen, but it is recommended to only use one.

• The random method is currently giving the best results.

After all options are chosen, press "Do Face Reconstruction" and wait.

• The frozen window is normal, the loading of the images can take up to 10 minutes, depending on the amount of images. Depending on your hardware the optimization can take a long time, the preview window should update every 30-60 seconds.