Unsupervised Depth Estimation, 3D Face Rotation and Replacement

TL;DR

This paper introduces DepthNets, an unsupervised method for estimating 3D facial structure and transformations from a single image, enabling face pose manipulation and re-synthesis without ground-truth depth data.

Contribution

It proposes a novel unsupervised framework for 3D face modeling and pose transformation using depth inference and differentiable loss, with additional adversarial techniques for head re-synthesis.

Findings

Successfully estimates 3D facial keypoints without ground-truth depth.

Enables face frontalization, pose transfer, and 3D modeling.

Uses adversarial image translation for head re-synthesis.

Abstract

We present an unsupervised approach for learning to estimate three dimensional (3D) facial structure from a single image while also predicting 3D viewpoint transformations that match a desired pose and facial geometry. We achieve this by inferring the depth of facial keypoints of an input image in an unsupervised manner, without using any form of ground-truth depth information. We show how it is possible to use these depths as intermediate computations within a new backpropable loss to predict the parameters of a 3D affine transformation matrix that maps inferred 3D keypoints of an input face to the corresponding 2D keypoints on a desired target facial geometry or pose. Our resulting approach, called DepthNets, can therefore be used to infer plausible 3D transformations from one face pose to another, allowing faces to be frontalized, transformed into 3D models or even warped to another pose and facial geometry. Lastly, we identify certain shortcomings with our formulation, and explore adversarial image translation techniques as a post-processing step to re-synthesize complete head shots for faces re-targeted to different poses or identities.