Home

Awesome

Controllable and Guided Face Synthesis for Unconstrained Face Recognition

European Conference on Computer Vision (ECCV 2022). [Arxiv, PDF, Supp, Project]

Feng Liu, Minchul Kim, Anil Jain, Xiaoming Liu

We propose a controllable face synthesis model (CFSM) that can mimic the distribution of target datasets in a style latent space. CFSM learns a linear subspace with orthogonal bases in the style latent space with precise control over the diversity and degree of synthesis. Furthermore, the pre-trained synthesis model can be guided by the face recognition (FR) model, making the resulting images more beneficial for FR model training. Besides, target dataset distributions are characterized by the learned orthogonal bases, which can be utilized to measure the distributional similarity among face datasets. Our approach yields significant performance gains on unconstrained benchmarks, such as IJB-B, IJB-C, TinyFace and IJB-S (+5.76% Rank1).

Introduction

This paper aims to answer the following three questions:

Prerequisites

This code is developed with

Stage1: Controllable Face Synthesis Model (CFSM)

Please refer to CFSM/README for the details.

Stage2: Guided Face Synthesis for Face Recognition

Please refer to GuidedFaceRecognition/README for the details.

Citation

@inproceedings{liu2022cfsm,
title={Controllable and Guided Face Synthesis for Unconstrained Face Recognition},
author={Liu, Feng and Kim, Minchul and Jain, Anil and Liu, Xiaoming},
booktitle={ECCV},
year={2022}}

Acknowledgments

Here are some great resources we benefit from:

License

MIT License

Contact

For questions feel free to post here or drop an email to - liufeng6@msu.edu