STAR Loss: Reducing Semantic Ambiguity in Facial Landmark Detection

06/05/2023
by   Zhenglin Zhou, et al.
0

Recently, deep learning-based facial landmark detection has achieved significant improvement. However, the semantic ambiguity problem degrades detection performance. Specifically, the semantic ambiguity causes inconsistent annotation and negatively affects the model's convergence, leading to worse accuracy and instability prediction. To solve this problem, we propose a Self-adapTive Ambiguity Reduction (STAR) loss by exploiting the properties of semantic ambiguity. We find that semantic ambiguity results in the anisotropic predicted distribution, which inspires us to use predicted distribution to represent semantic ambiguity. Based on this, we design the STAR loss that measures the anisotropism of the predicted distribution. Compared with the standard regression loss, STAR loss is encouraged to be small when the predicted distribution is anisotropic and thus adaptively mitigates the impact of semantic ambiguity. Moreover, we propose two kinds of eigenvalue restriction methods that could avoid both distribution's abnormal change and the model's premature convergence. Finally, the comprehensive experiments demonstrate that STAR loss outperforms the state-of-the-art methods on three benchmarks, i.e., COFW, 300W, and WFLW, with negligible computation overhead. Code is at https://github.com/ZhenglinZhou/STAR.

READ FULL TEXT

page 1

page 2

page 4

page 7

page 14

research
03/26/2019

Semantic Alignment: Finding Semantically Consistent Ground-truth for Facial Landmark Detection

Recently, deep learning based facial landmark detection has achieved gre...
research
03/31/2022

Adaptive Mean-Residue Loss for Robust Facial Age Estimation

Automated facial age estimation has diverse real-world applications in m...
research
03/21/2023

The Treasure Beneath Multiple Annotations: An Uncertainty-aware Edge Detector

Deep learning-based edge detectors heavily rely on pixel-wise labels whi...
research
09/21/2022

Uncertainty-aware Label Distribution Learning for Facial Expression Recognition

Despite significant progress over the past few years, ambiguity is still...
research
03/21/2023

PRISE: Demystifying Deep Lucas-Kanade with Strongly Star-Convex Constraints for Multimodel Image Alignment

The Lucas-Kanade (LK) method is a classic iterative homography estimatio...
research
08/25/2020

Confidence-aware Adversarial Learning for Self-supervised Semantic Matching

In this paper, we aim to address the challenging task of semantic matchi...
research
06/15/2018

Unsupervised Training for 3D Morphable Model Regression

We present a method for training a regression network from image pixels ...

Please sign up or login with your details

Forgot password? Click here to reset