Paper Abstract and Keywords |
Presentation |
2017-12-13 08:45
Improvement of Sign Language Recognition Performance by Integration of Multiple Classifiers and Feature Elements Tatsunori Ozawa, Yuna Okayasu (KAIT), Maitai Dahlan (Chulalongkorn University), Hiromitsu Nishimura, Hiroshi Tanaka (KAIT), Daisuke Kobayashi, Michio Iwamoto, Shuji Kato (KCC corp.) |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
The authors have been studying sign language recognition method using color gloves and optical cameras. We extracted 6 feature elements from the image data of color gloves, and we used HMM as recognition method by using these feature elements so far. In this paper, in addition to HMM that is based on probability of appearance, 5 other classifiers such as support vector machine, K neighborhood method, decision tree etc. were applied and their recognition performances was examined]. It was verified that the enhancement of recognition performance and stable recognition results can be obtained by unifying the results of each classifier. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Sign Language Recognition / Color gloves / Optical camera / Feature elements / Unification / / / |
Reference Info. |
IEICE Tech. Rep. |
Paper # |
|
Date of Issue |
|
ISSN |
|
Download PDF |
|