Fast-GANFIT: Generative Adversarial Network for High Fidelity 3D Face Reconstruction

TL;DR

This paper introduces Fast-GANFIT, a method combining GANs and deep learning to achieve high-fidelity 3D face reconstruction from single images, capturing high-frequency facial textures with improved accuracy.

Contribution

It presents a novel approach using GANs trained on large datasets for detailed facial texture prior, and a self-supervised regression method for robust 3D face fitting.

Findings

Achieves photorealistic 3D face reconstructions with high-frequency details.

Demonstrates robustness and efficiency in the fitting process.

Outperforms previous methods in texture quality and identity preservation.

Abstract

A lot of work has been done towards reconstructing the 3D facial structure from single images by capitalizing on the power of Deep Convolutional Neural Networks (DCNNs). In the recent works, the texture features either correspond to components of a linear texture space or are learned by auto-encoders directly from in-the-wild images. In all cases, the quality of the facial texture reconstruction is still not capable of modeling facial texture with high-frequency details. In this paper, we take a radically different approach and harness the power of Generative Adversarial Networks (GANs) and DCNNs in order to reconstruct the facial texture and shape from single images. That is, we utilize GANs to train a very powerful facial texture prior \edit{from a large-scale 3D texture dataset}. Then, we revisit the original 3D Morphable Models (3DMMs) fitting making use of non-linear optimization to find the optimal latent parameters that best reconstruct the test image but under a new perspective. In order to be robust towards initialisation and expedite the fitting process, we propose a novel self-supervised regression based approach. We demonstrate excellent results in photorealistic and identity preserving 3D face reconstructions and achieve for the first time, to the best of our knowledge, facial texture reconstruction with high-frequency details.