METHOD OF, SERVER FOR AND PROGRAM FOR DETECTING COGNITIVE AND SPEECH DISORDER BASED ON TIME-VARIANT VISUAL FEATURE OF FACE

• Main Authors: MORITZ WERNER AMADEUS EINFALT, MATTHEW LEE, WILCOX LYNN D, LYNDON KENNEDY, RAINER WOLFGANG LIENHART
• Format: Patent
• Language: English; Japanese
• Published: 13.08.2020
• Subjects: ACOUSTICS; CALCULATING; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; MUSICAL INSTRUMENTS; PHYSICS; SPEECH ANALYSIS OR SYNTHESIS; SPEECH OR AUDIO CODING OR DECODING; SPEECH OR VOICE PROCESSING; SPEECH RECOGNITION

To provide a system for, a method of, and a program for generating predictive information for use in prediction of a user state based on the input video with use of visual facial features of a user.SOLUTION: There is provided a computer-implemented method of evaluating a user state. The method includes the steps of: generating a landmark of a face on a received input video; defining points related to an interest regions of a user face; defining a conversation period based on the respective positions of the defined points; measuring, during the conversation period, frequency of rest, repeated patterns, and diversity of vocabulary; and gathering the measured results to generate predictive information for use in prediction of the user state.SELECTED DRAWING: Figure 5 【課題】ユーザの視覚的な顔の特徴を使用し、入力ビデオに基づいてユーザの状態を予測する予測情報を生成するシステム、方法、及びプログラムを提供する。【解決手段】ユーザの状態を評価するコンピュータ実装の方法が提供され、これは、受信した入力ビデオ上に顔のランドマークを生成し、ユーザの顔の関心領域に関連する点を定義し、定義された点のそれぞれの位置に基づいて会話の期間を定義し、会話の期間中に、休止頻度、繰り返しパターン、及び語彙の多様性を測定し、測定を集約して、ユーザの状態を予測する予測情報を生成すること、を含む。【選択図】図５