CapStARE: Capsule-based Spatiotemporal Architecture for Robust and Efficient Gaze Estimation

Miren Samaniego, Igor Rodriguez, Elena Lazkano
9/24/2025
cs.CV

Official code: toukapy/capsStare (9 stars, last commit 2025-09-30).

Abstract

We introduce CapStARE, a capsule-based spatio-temporal architecture for gaze estimation that integrates a ConvNeXt backbone, capsule formation with attention routing, and dual GRU decoders specialized for slow and rapid gaze dynamics. This modular design enables efficient part-whole reasoning and disentangled temporal modeling, achieving state-of-the-art performance on ETH-XGaze (3.36) and MPIIFaceGaze (2.65) while maintaining real-time inference (< 10 ms). The model also generalizes well to unconstrained conditions in Gaze360 (9.06) and human-robot interaction scenarios in RT-GENE (4.76), outperforming or matching existing methods with fewer parameters and greater interpretability. These results demonstrate that CapStARE offers a practical and robust solution for real-time gaze estimation in interactive systems. The related code and results for this article can be found on: https://github.com/toukapy/capsStare

View on arXivView PDF

Code Implementations(1)

toukapy/capsStareOfficial100%
90PythonDec 31, 20241 years ago

Similar Papers

Learning Spatio-Temporal Feature Representations for Video-Based Gaze Estimation

86% match2025

CrossGaze: A Strong Method for 3D Gaze Estimation in the Wild

84% match2024

GazeTarget360: Towards Gaze Target Estimation in 360-Degree for Robot Perception

84% match1 repo62025

GazeGen: Gaze-Driven User Interaction for Visual Content Generation

83% match2024

Inference-Time Gaze Refinement for Micro-Expression Recognition: Enhancing Event-Based Eye Tracking with Motion-Aware Post-Processing

83% match1 repo62025

Gaze-LLE: Gaze Target Estimation via Large-Scale Learned Encoders

83% match3 repos8442026

Cite this paper

@article{samaniego2025capstare,
  title  = {CapStARE: Capsule-based Spatiotemporal Architecture for Robust and Efficient Gaze Estimation},
  author = {Miren Samaniego and Igor Rodriguez and Elena Lazkano},
  year   = {2025},
  eprint = {2509.19936},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2509.19936v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2509.19936.svg)](https://codeofpaper.com/paper/2509.19936)

Discussion