Empowering Mainstreamed Deaf Students With Multiple Video Perspectives
By Kushalnagar, Raja S.; Jehan-Francois Paris; Technology and Deaf Education Symposium: Exploring Instructional and Access Technologies, June 21-23, 2010,Publication Date: 2010
Paper presents a personal multiple view perspective approach that enables deaf and hard of hearing (DHH) students to record multiple views of a classroom presentation on their laptops through the use of multiple web cameras. The Multiple View Perspectives (MVP) approach is a content-aware video encoding and delivery technique that reduces cognitive load and video processing demand by capturing and delivering multiple non-overlapping small targeted video streams. It ignores less interesting regions for further file space and video processing savings. MVP enables HDD students to see multiple perspectives simultaneously without losing essential information, by placing the views side by side. This universal access approach benefits other classroom participants including the presenter, other students, and interpreters or captioners. The feasibility of automated indexing and synchronization of captions with recorded lectures was explored through a comparison of the real-time captioner’s text file and the automated caption text file. This feasibility study revealed that the poorest classroom recording, captured at an angle of about 10 degrees from perpendicular off the caption video screen, yielded an accuracy rate of 70 percent, whereas one of the best recordings, captured a zero degrees from perpendicular, yielded an accuracy rate of 98 percent. Future work planned to improve the MVP system is discussed.
Published by: National Technical Institute for the Deaf (Website:http://www.ntid.rit.edu)
Link to text: http://www.rit.edu/~w-tecsym/cgi-bin/papers.cgi?T1B
