An Evaluation System for Voice Training by Image Processing

Abstract

Conversation is a fundamental way of communication. The ability in producing a speech sound is improved by talking in daily life. Nevertheless, there are people whose voice is hard to hear, and it is known that such voice can be cleared by training. In this report, we discuss the evaluation method by image processing on the voice training. Obtained face images are processed by spatial filter to detect the motion size and speed. The result shows that the ability of quantitative evaluation of this method in the condition of same-subject and fixed camera-face distance.