Study on multi-channel low-bitrate video transmission for sign languages under mobile communication
Grant-in-Aid for Scientific Research (C)
|Allocation Type||Single-year Grants|
|Research Institution||HIROSHIMA UNIVERSITY|
MIYAO Junichi Hiroshima University, Faculty of Integrated Arts and Sciences, Associate Professor, 総合科学部, 助教授 (30200124)
|Project Period (FY)
2002 – 2004
Completed(Fiscal Year 2004)
|Budget Amount *help
¥2,700,000 (Direct Cost : ¥2,700,000)
Fiscal Year 2004 : ¥1,000,000 (Direct Cost : ¥1,000,000)
Fiscal Year 2003 : ¥1,000,000 (Direct Cost : ¥1,000,000)
Fiscal Year 2002 : ¥700,000 (Direct Cost : ¥700,000)
|Keywords||sign language / video compression / low bitrate / mobile / Java / テレビ会議 / エージェント / マイグレーション|
In this research, low-bitrate sign language video communication for audio handicapped persons is studied in case of using mobile phones or notebook computers. The followings(1)-(6)are results of this research.
(1)Ultra low-bitrate sign language video communication method : It may be impossible to realize ultra low-bitrate video communication for sign languages by using common video compression methods. Thus, semantic frame extraction which identifies necessary and sufficient frames to recognize sign language, is proposed.
(2)Analysis of important elements in recognition of sign language : In order to identify important elements in representation, 14 words are analyzed. Then, it is showed that straight fingers with their direction and hand motion are important.
(3)Algorithm of semantic frame extraction : An algorithm is developed to extract semantically important frames according to above results.
(4)Software construction of the distributed system : As the proposed system becomes a distributed system, programming languages and software construction is considered. It is seemed that the language should be Java and a agent system should be used.
(5)Quantitative evaluation of a compression algorithm of a sign language : The proposed algorithm is evaluated. The results show that compared to MPEG2, amount of data is reduced to less than 50%, and optimization of the algorithm realizes 15% or more reduction. These results show effectiveness of this research.
(6)Obtained results are written in academic papers.
Research Products (13results)