The Speech Segmentation Module on the basis of Spectrum Variation Analysis

Abstract

Speech segmentation is intended for separation of homogeneous oscillation parts of the signal which correspond to different types of phonemes. The speech segmentation module is the fundamental element of speechlike signal generating systems for information security. The approach of the development of the speech segmentation module is considered.

Keywords

• speech segmentation
• phonetic unit
• speech classification
• voice activity detection
• digital signal processing
• spectral analysis
• signal cestrum