Domain Engineering for Applied Monocular Reconstruction of Parametric Faces

TL;DR

This paper introduces a novel machine learning approach for reconstructing parametric 3D face models from single images, leveraging synthetic data and domain adaptation to improve accuracy and training efficiency in industrial applications.

Contribution

It presents a new method for Face-to-Parameters reconstruction that combines synthetic data, domain decomposition, and domain adaptation, advancing the state-of-the-art in monocular face reconstruction.

Findings

Improved accuracy in face parameter reconstruction from single images.

Enhanced training efficiency for parametric face models.

Open-source code enables quantitative evaluation and practical industrial use.

Abstract

Many modern online 3D applications and video games rely on parametric models of human faces for creating believable avatars. However, manually reproducing someone's facial likeness with a parametric model is difficult and time-consuming. Machine Learning solution for that task is highly desirable but is also challenging. The paper proposes a novel approach to the so-called Face-to-Parameters problem (F2P for short), aiming to reconstruct a parametric face from a single image. The proposed method utilizes synthetic data, domain decomposition, and domain adaptation to address multifaceted challenges in solving the F2P. The open-sourced codebase illustrates our key observations and provides means for quantitative evaluation. The presented approach proves practical in an industrial application; it improves accuracy and allows for more efficient models training. The techniques have the potential to extend to other types of parametric models.