Simplified Eye Gesture Recognition for control of Assistive Devices

DESCRIPTION (provided by applicant): This project aims to provide a simpler, more user-friendly system for eye-controlled communication and computer interaction. Although eye-tracking systems are currently available, the existing systems tend to be very expensive, typically costing thousands of dollars, and most importantly, are difficult to set up, calibrate, and use. In this project, we will develop an eye-based communication device that is much simpler and is easy to use. It will be based on detecting qualitative eye gestures for communication, rather than on precisely measuring eye position. In the future, this device holds promise to integrate well with advanced software we have identified for text generation, word selection, and disambiguation. In this project, we will first identify practical eye gestures with a consortium of practitioners, technologists, and users. We will then develop a proof-of-concept I-gesture device using low-cost optical and electronic components. Java-based software will be developed for the PC to enable realistic testing of the device. The proof-of-concept device will be tried by a group of volunteer participants with disabilities to show feasibility of the I-gesture technology, and preliminary cost estimates will be derived to help predict user acceptance of the device. More than one million people in the United States have communications disorders severe enough to require augmentative communications technology, and many tens of thousands have disabilities so severe that few alternatives other than eye-based communication is possible. However, existing eye tracker solutions tend to be difficult to use for many individuals. This project will result in a device that is much more practical for a significant proportion of this population, and also promises to be lighter, smaller, and competitively priced.