@techreport{oai:ipsj.ixsq.nii.ac.jp:00232990, author = {工藤, 達郎 and 長谷川, 幸代 and 田尻, 姿穂 and 藤原, 百合 and 山本, 一郎 and 工藤, 孔梨子 and Tatsuro, Kudo and Sachiyo, Hasegawa and Shiho, Tajiri and Yuri, Fujiwara and Ichiro, Yamamoto and Kuriko, Kudo}, issue = {17}, month = {Mar}, note = {構音訓練において,発声時の音声器官の形状や動きを理解することは重要である.既存の教材は主に 2 次元画像が使用されており,口腔内の 3 次元表現には限界があった.そこで本研究では MetaHuman Technology(Epic Games, Inc)を用いて口腔内の形状や動作を可視化する試みと評価を行った.対象とした音声は構音訓練の基本となる [タ],[テ],[ト] である.健常者 10 名が発音する様子を前方と斜めから撮影し,映像データを参照しながら Unreal Engine(v.4.72,v5.2.1)の Control Rig と Sequencer の機能を用いて音声器官の 3DCG アニメーションを作成した.その後「口腔内のみ」「舌のみ」など表示形式を自由に選択しながら,任意の視点から観察できる可視化環境を構築した.構築したシステムは,口腔内の音声器官の形状や動きを理解する能力について専門家 4 名へのインタビューにより評価された.舌の動きや空気の流れが立体的に可視化され分かりやすいと評価されたが,舌の端のみを上げる動作など,構音訓練において重要だが既存パラメータでは再現できない舌の形状も明らかとなった.今後もデータ取得法や舌のモデルを改良し,高品質な可視化システムの開発に取り組む., It is important to understand the complex shape and movement of speech organs in the oral cavity for speech therapy. However, most existing study materials are two-dimensional, affording only a limited representation of complex shape and movement of speech organs in the oral cavity. In recent years, MetaHuman Technology (Epic Games, Inc), a platform that can present a realistic human body in real-time by adjusting body parameters, was recently developed. In this study, we report our attempt to visualize the oral cavity using MetaHuman as part of articulation intervention. The target speech sounds were [ta], [te], and [to], which form the basis of speech therapy. First, video recordings were taken of ten Japanese-speaking adults without speech and hearing problems. Using the typical speakers as a guide, we applied the Control Rig and Sequencer features of the Unreal Engine (v.4.72, v5.2.1) to generate computer graphic animations of MetaHuman's mouth. In addition, we have created a user interface that allows observation from any viewpoint while freely selecting display formats such as “oral cavity only” and “tongue only”. The constructed system was evaluated by the four experts in terms of its ability to understand the shape and movement of speech organs in the oral cavity. All experts evaluated the system favorably in terms of “ability to observe tongue movement from various angles,” “clear representation of airflow,” and “usefulness in clinical practice with magnified image of the face and translucent oral cavity.” However, there are some tongue shapes that are important in articulation training but cannot be reproduced with existing parameters, such as bilateral bracing and central groove formation. We will continue refining high-quality visualization systems for speech therapy.}, title = {MetaHumanを用いた構音訓練用口腔内ビジュアライザの試作と評価}, year = {2024} }