Dense Face Alignment

TL;DR

This paper introduces a dense 3D face alignment method using CNNs that fits face contours and features, outperforming existing landmark detection techniques on large-pose images in real-time.

Contribution

It presents the first approach for dense 3D face alignment that integrates contour and feature fitting, addressing dataset annotation inconsistencies.

Findings

Achieves high-quality dense 3D face fitting.

Outperforms state-of-the-art landmark detection methods.

Operates in real-time during testing.

Abstract

Face alignment is a classic problem in the computer vision field. Previous works mostly focus on sparse alignment with a limited number of facial landmark points, i.e., facial landmark detection. In this paper, for the first time, we aim at providing a very dense 3D alignment for large-pose face images. To achieve this, we train a CNN to estimate the 3D face shape, which not only aligns limited facial landmarks but also fits face contours and SIFT feature points. Moreover, we also address the bottleneck of training CNN with multiple datasets, due to different landmark markups on different datasets, such as 5, 34, 68. Experimental results show our method not only provides high-quality, dense 3D face fitting but also outperforms the state-of-the-art facial landmark detection methods on the challenging datasets. Our model can run at real time during testing.