Development of communication support system based on intent recognition for cerebral palsy

2016 Fiscal Year Final Research Report

• Project/Area Number: 25282053
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Partial Multi-year Fund
• Section: 一般
• Research Field: Educational technology
• Research Institution: Kobe University
• Principal Investigator: Takiguchi Tetsuya 神戸大学, 都市安全研究センター, 准教授 (40397815)
• Co-Investigator(Kenkyū-buntansha): 有木 康雄 神戸大学, 学内共同利用施設等, 教授 (10135519) ; 高田 哲 神戸大学, 保健学研究科, 教授 (10216658) ; 中井 靖 宮崎大学, 教育文化学部, 講師 (80462050) ; 榎並 直子 神戸大学, 学内共同利用施設等, 助教 (80628925) ; 中川 誠司 国立研究開発法人産業技術総合研究所, その他部局等, 研究員 (70357614)
• Project Period (FY): 2013-04-01 – 2017-03-31
• Keywords: ヒューマン・インターフェース

Outline of Final Research Achievements

An utterance style of a person with an articulation disorder, such as cerebral palsy, is different from that of physically unimpaired persons, and his/her utterance is often unstable or unclear, which makes it difficult for them to communicate. To develop a communication-support system for them, we propose a new automatic speech recognition (ASR) system using a new acoustic feature extraction technique, a voice conversion (VC) method for articulation disorders that converts unclear utterances to clear utterances, and a multi-modal utterance recognition system using a novel feature integration technique based on a machine-learning approach. Experimental results demonstrated that our ASR, VC, and multimodal recognition methods could improve the speech recognition accuracy, the listening speech quality, and the multimodal (speech and image) recognition accuracy in comparison with conventional approaches, respectively.

Free Research Field

メディア情報処理