This is a revised proposal to further develop and evaluate a display-reader for blind users, a device that reads a loud the LED and LCD displays found on many electronic appliances and instruments. Such a display-reader would be indispensable to a great number of blind persons, as more and more ordinary devices has such displays (ovens, washers, coffee makers, fax machines, telephones, VCRs, thermostats). The key challenges are in developing the machine vision algorithms to read a broad range of displays, investigating the human factors and hardware engineering issues that go into image acquisition and useful audio or other output, and incorporating these findings into a practical display-reader. The device will have two embodiments: variants of the current prototype, based on a portable PC or palmtop computer and a consumer CCD camera; ad a completely self-contained hand-held version. The hardware for the latter will be developed based on one of the new, high-speed digital signal processor chips. Criteria for a successful prototype are: 1)speed: one second to read a display, 2)accuracy: 98 percent of display outputs read correctly in trials wit blind subjects, and 3)ease of use: handling, user-interface, battery life, and other subjective measures obtained through blind user feedback during trials.

Proposed Commercial Applications

Not Available

Agency
National Institute of Health (NIH)
Institute
National Eye Institute (NEI)
Type
Small Business Innovation Research Grants (SBIR) - Phase II (R44)
Project #
2R44EY011827-02
Application #
6299098
Study Section
Special Emphasis Panel (ZRG1-SSS-R (02))
Program Officer
Oberdorfer, Michael
Project Start
1997-09-30
Project End
2003-02-28
Budget Start
2001-03-05
Budget End
2002-02-28
Support Year
2
Fiscal Year
2001
Total Cost
$555,256
Indirect Cost
Name
Blindsight Corporation
Department
Type
DUNS #
City
Berkeley
State
CA
Country
United States
Zip Code
94704