An intelligent computing system to detect material

Abstract

The detection of material with high accuracy is a hot research area in material science nowadays. Although, there is significant advancement in modeling new material but there is uncertainty in bringing new material. An intelligent computing system using machine learning algorithms plays an important role in the detection and discovery of new material from concept to reality. The machine learning algorithm can be applied from design phase to material synthesis phase and also in development and validation of new material. This paper deals with image processing, machine learning and material science concept that affect properties, research, problems, challenges and probable solutions to detect and discover material. Different types of filters are utilized for detection of new material. Further, features of materials are extracted by using histogram, wavelet transform and color moments. The k-Nearest Neighbor (k-NN) and multi-Support Vector Machine (SVM) machine learning algorithms are utilized to classify the materials and their performances are also compared. The accuracy obtained from multi-SVM outperforms k-NN.