Facial Expression Research Group 3D Database

Expression transfer from humans to stylized character using our work. Primary character: Mery, secondary characters (left to right): Bonnie, Ray and Malcolm. We show transfer of expressions belonging to 6 categories as in FERG-DB.


Facial Expression Research Group 3D Database (FERG-3D-DB) is a database of 3D rigs of stylized characters with annotated facial expressions. The database contains 39574 annotated examples for four stylized characters (2 females and 2 males) - Mery, Bonnie, Ray and Malcolm. Each example is a list of rig parameter values which when transferred to the 3D rig creates a particular facial expression. The expressions for each character are grouped into seven categories - anger, disgust, fear, joy, neutral, sadness and surprise.

The characters were modeled using the MAYA software. The 3D rigs of these characters can be obtained from:

The database consists of:
  • CSV files, one per character, combined in a password-protected zip file with a README about the details
    • The first row of each CSV file contains the names of the facial rig parameters to be manipulated to create an expression
    • Each of the following rows contains the values of the parameters for a particular facial expression
    • The first column contains the names of the examples. The expression category is mentioned in the names. Note, the names do not correspond to the image names in FERG-DB.
  • Python code to read the entries from the CSV files and transfer them to the 3D rigs to animate them



If you use this dataset, please cite our paper:

	author={D. {Aneja} and B. {Chaudhuri} and A. {Colburn} and G. {Faigin} and L. {Shapiro} and B. {Mones}}, 
	booktitle={2018 IEEE Winter Conference on Applications of Computer Vision (WACV)}, 
	title={Learning to Generate 3D Stylized Character Expressions from Humans}, 


Database access

The python code to generate facial expressions on the 3D MAYA rigs is available here on github. Please fill out this agreement form to get access to the zip file containing CSV files.



We would like to thank creators of the rigs we used in our project and Jamie Austad for creating the facial expressions on the rigs.