Terahertz spectroscopic material identification using approximate entropy and deep neural network

Abstract

Terahertz spectroscopy and imaging are a rapidly developed technique with important applications in many areas, such as medical imaging, security, chemistry, biochemistry, astronomy, communications, and manufacturing, to name a few. However, terahertz spectroscopy and imaging produce excessively high dimensional data which is prohibitive for common methods developed in the area of image processing. In this paper, we report our recent study on a novel classifier based on feature extraction using approximate entropy (ApEn). The classifier is initiated by analyzing the complexity of the terahertz spectrum, which is then combined with a deep neural network for material classification. Experimental results show that approximate entropy based features have high sensitive for detecting metal matrix composites, the accuracy of identification is up to 96.3%. Related algorithms for ApEn feature extraction and material classification are illustrated. An optimal parameter-embedding dimension, subject to classification accuracy for ApEn is studied.