A Video Quality Control Method Using Stereophonic Speech On TV Conference

Abstract

In this paper, a quality enhancement method is proposed for efficient coding of video in teleconference systems. The method precisely encodes the talker's face area without increasing average encoded bit rate. To specify the talker's face, delay value is calculated from stereophonic speeches by estimating cross-channel transfer function. Computer simulation results showed that the quality enhancement was well controlled by dynamically tracking the talker's face location using stereophonic speeches.