3D Facial Geometry Recovery from a Depth View with Attention Guided Generative Adversarial Network

TL;DR

This paper introduces AGGAN, a novel attention-guided GAN that reconstructs complete 3D facial geometry from a single depth view, outperforming existing methods in accuracy and robustness.

Contribution

The paper proposes a new AGGAN model that recovers full 3D facial geometry from one depth view using attention mechanisms and multiple loss functions, advancing beyond multi-view requirements.

Findings

AGGAN produces more complete 3D facial shapes.

It handles wider view angles and noise better.

Qualitative and quantitative results validate its effectiveness.

Abstract

We present to recover the complete 3D facial geometry from a single depth view by proposing an Attention Guided Generative Adversarial Networks (AGGAN). In contrast to existing work which normally requires two or more depth views to recover a full 3D facial geometry, the proposed AGGAN is able to generate a dense 3D voxel grid of the face from a single unconstrained depth view. Specifically, AGGAN encodes the 3D facial geometry within a voxel space and utilizes an attention-guided GAN to model the illposed 2.5D depth-3D mapping. Multiple loss functions, which enforce the 3D facial geometry consistency, together with a prior distribution of facial surface points in voxel space are incorporated to guide the training process. Both qualitative and quantitative comparisons show that AGGAN recovers a more complete and smoother 3D facial shape, with the capability to handle a much wider range of view angles and resist to noise in the depth view than conventional methods