01-10-2012, 11:15 AM
Eye Gaze Tracking for Human Computer Interaction
Eye Gaze Tracking.pdf (Size: 3.54 MB / Downloads: 60)
Introduction
With the invention of the computer in the middle of the last century there was also the need of an interface for users. In the beginning experts used teletype to interface with the computer. Due to the tremendous progress in computer technology in the last decades, the capabilities of computers increased enormously and working with a computer became a normal activity for nearly everybody. With all the possibilities a computer can offer, humans and their interaction with computers are now a limiting factor. This gave rise to a lot of research in the field of HCI (human computer interaction) aiming to make interaction easier, more intuitive, and more efficient. Interaction with computers is not limited to keyboards and printers anymore. Different kinds of pointing devices, touch-sensitive surfaces, high-resolution displays, microphones, and speakers are normal devices for computer interaction nowadays. There are new modalities for computer interaction like speech interaction, input by gestures or by tangible objects with sensors. A further input modality is eye gaze which nowadays finds its application in accessibility systems. Such systems typically use eye gaze as the sole input, but outside the field of accessibility eye gaze can be combined with any other input modality. Therefore, eye gaze could serve as an interaction method beyond the field of accessibility. The aim of this work is to find new forms of interactions utilizing eye gaze and suitable for standard users.
Motivation
Nowadays most eye-tracking systems work on video-based pupil detection and a reflection of an infrared LED. Video cameras became cheap over the last years and the price for a LED is negligible. Many computer devices come already with built-in cameras, such as mobile phones, laptops, and displays. Processor power still increases steadily and standard processors are powerful enough to process the video stream necessary to do eye tracking, at least on desktop and laptop computers. Head-tracking systems, which are necessary to give the users the freedom to move in front of their display, are also video-based. Such systems can be implemented unobtrusively with a second camera. If produced for the mass market, a future standard eye tracker should not cost much more than an optical mouse or a webcam today.
Eye Gaze and Human Communication
When we ask “What is this?” we often reference the object with the direction of our gaze and we assume that the person asked knows which object we are looking at. Bolt, a pioneer of gaze interaction research, expressed it this way:
‘Consider the case of asking the question “What is your favorite sport?” in the presence of several people, but looking at Mary, say, not at Frank, Judy, or Dave. Then you utter the selfsame words, but now looking at Dave. The question is a different question; the difference lies not in its verbal component, but in its intended addressee as given by eye.’ [Bolt 1982].
When we talk to somebody we normally address the person by looking at her or him, but staring at somebody without reason is not polite and should be avoided. These small examples on how we use the eyes for communication illustrate that our eyes move not only to enable vision. We are aware that others are aware where we are looking to and for this reason we need to have our eyes under control.
When comparing human with animal eyes the white in the human eyes is very distinctive (see Figure 1). In animal eyes, eyes of mammals in particular, which work in a similar way as human eyes, no white eyeball is visible. For this reason it is more difficult to determine the direction of an animal’s gaze than of a human’s. What role the awareness of the gaze’s direction played in the evolution of our species is a field for speculation. What is certain is that we use our eyes for communication. Eibl-Eibesfeldt discusses the topic [Eibl-Eibesfeldt 2004, p. 621] and the meaning of eye contact in general. He observed that looking at somebody opens a communication channel, but all ethnics and cultures interpret staring at somebody as an act of aggression.
Eye Gaze for Computer Input
An eye-gaze interface seems to be a promising candidate for a new interface technique, which may be more convenient than the ones we use. Traditionally, disabled people who cannot move anything except their eyes use eye gaze interaction. These systems are designed to direct the computer solely by the eyes. Such systems work well and are a great help for people who need them, but for others they are cumbersome and less efficient than keyboard and mouse. This contradicts the fact that looking is an easy task and that eye movements are fast. Consequently, eye-gaze interfaces for the masses need a different design to bring benefit to the average user.
Methods and Approach
The methods and approach used in this thesis are mostly experimental which is uncommon in computer science but appropriate in the field of HCI.
An iterative concept helps to get systematically closer to the development of new interaction methods based on gaze. Starting with a problem to solve the first step is to create ideas how to solve the problem. The resulting task is to realize the idea by creating a prototype. Typically, a pilot study gives the first hints for a redesign of the prototype and for the design of the user study. The data collected in the user study are the basis for an evaluation using statistical methods. Interviews with test users are the method to find out soft factors due to human aspects. Feeling comfortable with an interface or the fascination invoking commands by eye gaze are hard to measure by physical data. The results of a user study lead to further enhancements of the prototype and a follow-up user study. The iteration of modifying prototypes and conducting user studies leads to the final goal of understanding the implications of the tested prototype.
All prototypes used are based on an eye tracker and it’s API (application programming interface). Each prototype uses software especially developed for the purpose. Additionally built hardware prototypes, such as touch-sensitive mouse devices, allow to research multimodal input.
Contributions
Within an overview of the current state of research and technology this work discusses the accuracy of eye trackers and the need of a calibration process. The so-called calibration-free eye-tracking methods allow a simpler calibration process, which may not even be noticed by the user, but these methods still require calibration because of anatomic variability of individuals.
A further discussion focuses on theoretical aspects like Fitts’ law and the related speed and accuracy issues. The discussion shows that there is no reason applying Fitts’ law to eye movements. A general discussion of Fitts’ law was published in [Drewes 2010].
The chapter on eye gaze pointing presents a touch-sensitive mouse as novel input device. Based on the observation that people have problems to locate the mouse pointer, the idea is to position the mouse pointer at the gaze position when touching the mouse. The approach leads to an improvement of Zhai’s MAGIC (Mouse And Gaze Input Cascaded) pointing principle avoiding compensation methods needed for the MAGIC principle and giving the user a higher degree of intentional control. The core result of the chapter is that the use of gaze-assisted pointing saves an enormous amount of mouse movements.
Overview and Related Work
There is an immense knowledge on eye tracking and related fields such as the anatomy and physiology of the eye, the movement of the eyes and visual perception. The knowledge spreads over many disciplines of science including biology, medicine, psychology, and neurology. Even a rough overview on the topic could fill books. See Duchowski’s textbook on eye-tracking methodology [Duchowski 2002] for a more detailed overview.
This overview starts with a definition of eye tracking followed by a short history of eye tracking and the fields of application for eye trackers. The next two sections explain the eye-tracking technology and present available eye-tracker systems. A further section explains the commercial eye tracker used for all experiments and user studies in this work. Finally, the last section mentions the scientific work on eye tracking for interaction done in the past, but only on a general level as the following chapters have their own related work sections.