Audio and Video Speech Processing

Abstract Audio and video speech processing has been a trending research area in the recent years. Conversion of audio cues and video cues into text is really helpful for people having speech and hearing impairments. In this paper, we present the method of converting audio and video frames into the text which represents the corresponding spoken word. We have used a k-NN classifier to classify different speakers utterances, MFCC algorithm for audio classification as well as content-based retrieval and Support Vector Machine, a supervised learning method which generates input-output mapping functions from a set of labeled training data.