LOTR: Face Landmark Localization Using Localization Transformer

TL;DR

This paper introduces LOTR, a Transformer-based model for facial landmark localization that directly predicts coordinates, utilizing a novel smooth-Wing loss for improved training, and demonstrates superior performance on benchmark datasets.

Contribution

The paper proposes a new Transformer-based framework for facial landmark localization with a direct coordinate regression approach and a novel smooth-Wing loss function, enhancing accuracy and convergence.

Findings

LOTR outperforms existing methods on JD landmark dataset.

LOTR achieves promising results on WFLW dataset.

Using LOTR for face alignment improves face recognition performance.

Abstract

This paper presents a novel Transformer-based facial landmark localization network named Localization Transformer (LOTR). The proposed framework is a direct coordinate regression approach leveraging a Transformer network to better utilize the spatial information in the feature map. An LOTR model consists of three main modules: 1) a visual backbone that converts an input image into a feature map, 2) a Transformer module that improves the feature representation from the visual backbone, and 3) a landmark prediction head that directly predicts the landmark coordinates from the Transformer's representation. Given cropped-and-aligned face images, the proposed LOTR can be trained end-to-end without requiring any post-processing steps. This paper also introduces the smooth-Wing loss function, which addresses the gradient discontinuity of the Wing loss, leading to better convergence than standard loss functions such as L1, L2, and Wing loss. Experimental results on the JD landmark dataset provided by the First Grand Challenge of 106-Point Facial Landmark Localization indicate the superiority of LOTR over the existing methods on the leaderboard and two recent heatmap-based approaches. On the WFLW dataset, the proposed LOTR framework demonstrates promising results compared with several state-of-the-art methods. Additionally, we report the improvement in state-of-the-art face recognition performance when using our proposed LOTRs for face alignment.