一側性難聴者は音の定位や音声の認識が困難である.またノイズキャンセリングイヤホン等の機器の利用者は,周辺の音声を認識できない場合もある.そこで,本研究では,AR グラスと 360°マイクアレイを用いて, 音声到来方向と音声の文字起こし結果を表示するシステムを提案し,そのプロトタイプシステムを作成した.本システムでは,ネットワーク内のエッジサーバ内で音声認識処理を行うことで音声録音から AR グラスへの表示時間を短縮し,リアルタイムな音声支援を実現した.
(英)
People with unilateral hearing loss struggle with sound localization and voice recognition. Additionally, users of noise-cancelling earphones and similar devices may also find it difficult to recognize surrounding sounds. To address these issues, this study proposes a system that uses AR glasses and a 360-degree microphone array to display both the direction of incoming voices and transcriptions of the voices. We have developed a prototype system to demonstrate this. In this system, voice recognition processing is carried out on edge servers within the network, reducing the time from voice recording to display on the AR glasses and achieving real-time voice assistance.