Description / 描述

This is a simple demonstration of the Proof-of-Concept (POC) webcam gaze tracking technology that is intended to be used in the web-based eye monitoring software for users who are suffering from Age-Related Macular Degeneration (AMD).

The technology basically works by detecting the position of the eye landmarks and iris centres of the users from webcam images, and use the distances between these landmarks as features which are then fed to a machine learning model that will learn how these distances can be translated to where the user is looking at on a screen (calibration), and attempt to predict the user's gaze location on the screen from the webcam images once learning is completed. This technology is still under development, so the performance from the demostration still has a large room for improvement.

这是一个简单的摄像头视线追踪技术的概念验证（POC）演示, 旨在应用于眼部监测软件以便监测患有与年龄相关的黄斑变性（AMD）的用户的病情。 该技术的基本工作原理是通过从摄像头图像中检测用户的眼部标志和虹膜中心的位置，并将这些标志之间的距离作为特征而输入到机器学习模型中。 该模型将学习如何将这些距离转化为用户在屏幕上的注视点（校准过程），并尝试从摄像头图像中预测用户在屏幕的注视位置。 这项技术仍在开发中，所以演示中的表现还是有很大的进步空间。