Classification of Stuttering Dysfluency Using Enhanced Feature Selection Techniques

Abstract

One of the main speech fluency disorder is Stuttering or Stammering, which is identified by various characteristics like prolongation, repetitions of words and syllables, frequent pause, incomplete words and revisions. Stuttering is a complex disorder that may encompass social and emotional elements. An automatic stutter dysfluency recognition system is important to identify the problem in early stages so that therapy can be given to improve their speech communication. Automatic recognition consists of three main steps, namely, segmentation, feature extraction and classification. This research work focuses on proposing techniques that can enhance the recognition of dysfrequencies in stuttered speech signals, without manual intervention. In particular, the research work is involved in the identification of methods that can improve the process of automatic identification of dysfluencies in recorded stuttered speech signals. newline