Local binary patterns for 1-D signal processing

Abstract

Local Binary Patterns (LBP) have been used in 2-D image processing for applications such as texture segmentation and feature detection. In this paper a new 1-dimensional local binary pattern (LBP) signal processing method is presented. Speech systems such as hearing aids require fast and computationally inexpensive signal processing. The practical use of LBP based speech processing is demonstrated on two signal processing problems: - (i) signal segmentation and (ii) voice activity detection (VAD). Both applications use the underlying features extracted from the 1-D LBP. The proposed VAD algorithm demonstrates the simplicity of 1-D LBP processing with low computational complexity. It is also shown that distinct LBP features are obtained to identify the voiced and the unvoiced components of speech signals.