浏览器内轻量眼动追踪，仅需9次校准即可高精度适配新用户。
WEBEYETRACK: Scalable Eye-Tracking for the Browser via On-Device Few-Shot Personalization
- 在浏览器中部署轻量级模型，结合头姿估计与设备端少样本学习。
- 在GazeCapture数据集上误差仅2.32厘米，iPhone 14上推理速度达2.4毫秒。
- 适合隐私敏感场景，支持快速个性化，开源可复现。
随着AI进步，新型注视估计方法虽超越现有最先进（SOTA）基准，但其实际应用仍不及商业眼动追踪方案。模型体积、推理时延和隐私问题常被忽视。而基于网络摄像头的追踪方法因头部移动导致精度不足。为此，我们提出WebEyeTrack框架，将轻量级SOTA注视估计模型直接集成于浏览器中。该框架结合模型化头姿估计与设备端少样本学习，仅需九个校准样本（k < 9）即可适配新用户。在GazeCapture数据集上实现2.32厘米误差，iPhone 14上实现实时推理速度2.4毫秒。代码已开源：https://github.com/RedForestAi/WebEyeTrack。
原文摘要 · Abstract (English)
With advancements in AI, new gaze estimation methods are exceeding state-of-the-art (SOTA) benchmarks, but their real-world application reveals a gap with commercial eye-tracking solutions. Factors like model size, inference time, and privacy often go unaddressed. Meanwhile, webcam-based eye-tracking methods lack sufficient accuracy, in particular due to head movement. To tackle these issues, we introduce We bEyeTrack, a framework that integrates lightweight SOTA gaze estimation models directly in the browser. It incorporates model-based head pose estimation and on-device few-shot learning with as few as nine calibration samples (k < 9). WebEyeTrack adapts to new users, achieving SOTA performance with an error margin of 2.32 cm on GazeCapture and real-time inference speeds of 2.4 milliseconds on an iPhone 14. Our open-source code is available at https://github.com/RedForestAi/WebEyeTrack.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。