Three-Dimensional Gaze Estimation for Single Camera Systems

内容简介

On the basis of introducing some theories of gaze tracking and helping to understand the essence of gaze tracking, this book presents new methodologies that can estimate the 3-D gaze using a single-camera system. Regardless of whether the system is configured with one or two light sources, and the type of light source is visible or infrared, there are corresponding methods to estimate the 3-D line-of-sight and point-of-regard. The feasibility validation and error source analysis in simulations, as well as the accuracy evaluation and performance analysis in practical experiments, elucidate the characteristics of these methods.

目录

Chapter 1 Introduction

1.1 Background and motivation

1.2 Outline of the book

Chapter 2 Preliminaries

2.1 Basic structure of the eyeball

2.2 Gaze-related eyeball information

2.3 Eye movement forms

2.4 Visual information used for gaze tracking

2.5 Classification of gaze tracking systems

Chapter 3 System Calibration for Gaze Tracking Systems

3.1 Camera calibration

3.2 Global calibration of light sources and screen

3.3 Ball-assisted calibration of light sources and screen

Chapter 4 Image Processing for Gaze Tracking Systems

4.1 Eye region detection

4.2 Iris detection

4.3 Pupil detection

4.4 Glint localization

Chapter 5 3-D Gaze Estimation Based on Iris Radius Calibration for Systems with Single Camera and Single Light Source

5.1 Algorithms

5.2 Simulations

5.3 Practical experiments

Chapter 6 3-D Gaze Estimation Based on Corneal Radius Calibration for Systems with Single Camera and Single Light Source

6.1 Algorithms

6.2 Simulations

6.3 Practical experiments

Chapter 7 3-D Gaze Estimation for Systems with Single Camera and Two Collinear Light Sources

7.1 Algorithms

7.2 Simulations

7.3 Practical experiments

Chapter 8 3-D Gaze Estimation Based on Cross-Ratio Using Weighted Average and Polynomial Compensation

8.1 Algorithms

8.2 Simulations

8.3 Practical experiments

References