Prototype of a support person-assisted machine learning system for the communicative acquisition potential of children with severe multiple disabilities.

2022 Fiscal Year Final Research Report

• Project/Area Number: 19K11417
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 59010:Rehabilitation science-related
• Research Institution: Kagawa University
• Principal Investigator: Miyazaki Eiichi 香川大学, 教育学部, 教授 (30253248)
• Co-Investigator(Kenkyū-buntansha): 坂井 聡 香川大学, 教育学部, 教授 (90403766)
• Project Period (FY): 2019-04-01 – 2023-03-31
• Keywords: 障害者支援 / 重度重複障害児 / モーションパターン / ディープラーニング / ＩＣＴ

Outline of Final Research Achievements

Communication with children with severe multiple disabilities, who have multiple disabilities, is often difficult even for supporters with many years of experience. Therefore, we have developed a communication support interface that uses deep learning to discriminate the intentionality of children with severe multiple disabilities based on their behavior patterns by capturing their interactions with a web camera in advance. The use of deep learning has brought about the possibility of feature extraction, which had been difficult to discriminate, and has given inexperienced supporters clues for discriminating intentionality. This understanding of the surroundings made the disabled children aware of their own communication potential, which encouraged them to communicate spontaneously, and brought about the possibility of improving the quality of their daily life at school and at home.

Free Research Field

リハビリテーション科学関連

Academic Significance and Societal Importance of the Research Achievements

重度重複障害児は、発声や発音に重い障害があり、筆記などの表現手段も利用できないことが多く、周囲に自分の意志や要求の伝達が困難であった。その結果、受動的なコミュニケーションが主となり、根本的なコミュニケショーション能力不足を招いていた。これが重度重複障害児の学習や精神的な発達を妨げており、意思を表出できるコミュニケーション方法の開発が教育現場だけに止まらず、多くの生活の場面において望まれていた。
本研究では意図性判別の支援可能な入力インタフェースを開発し、重度重複障害児に自身が潜在的に持っているコミュニケーション能力に気づきを与え、日常生活において受動的態度から能動的態度を涵養する可能性を導いた。