Improving Domain Generalization on Gaze Estimation via Branch-out Auxiliary Regularization
Preprint
Abstract
Despite remarkable advancements, mainstream gaze estimation techniques, particularly appearance-based methods, often suffer from performance degradation in uncontrolled environments due to variations in illumination and individual facial attributes. Existing domain adaptation strategies, limited by their need for target domain samples, may fall short in real-world applications. This letter introduces Branch-out Auxiliary Regularization (BAR), an innovative method designed to boost gaze estimation's generalization capabilities without requiring direct access to target domain data. Specifically, BAR integrates two auxiliary consistency regularization branches: one that uses augmented samples to counteract environmental variations, and another that aligns gaze directions with positive source domain samples to encourage the learning of consistent gaze features. These auxiliary pathways strengthen the core network and are integrated in a smooth, plug-and-play manner, facilitating easy adaptation to various other models. Comprehensive experimental evaluations on four cross-dataset tasks demonstrate the superiority of our approach.Keywords
All Related Versions
This publication has 8 references indexed in Scilit:
- PnP-GA+: Plug-and-Play Domain Adaptation for Gaze Estimation Using Model VariantsIEEE Transactions on Pattern Analysis and Machine Intelligence, 2024
- End-to-End Video Gaze Estimation via Capturing Head-Face-Eye Spatial-Temporal Interaction ContextIEEE Signal Processing Letters, 2023
- A Driver Gaze Estimation Method Based on Deep LearningSensors, 2022
- Gaze Estimation by Exploring Two-Eye AsymmetryIEEE Transactions on Image Processing, 2020
- MPIIGaze: Real-World Dataset and Deep Appearance-Based Gaze EstimationIEEE Transactions on Pattern Analysis and Machine Intelligence, 2017
- Combining Head Pose and Eye Location Information for Gaze EstimationIEEE Transactions on Image Processing, 2011
- A novel non-intrusive eye gaze estimation using cross-ratio under large head motionComputer Vision and Image Understanding, 2005
- Eye movements in reading and information processing: 20 years of research.Psychological Bulletin, 1998