Local Binary Patterns and Extreme Learning Machine based Texture Classification of Marbles

Abstract

Separation of marbles into similar textures is a demanding and specialized process. In general, this process is carried out by experts. Marble plates with different textures to be placed in the same group is not a desirable status. This situation may lead to customer dissatisfaction in the sales process and post-sales. In this study, we propose a marble classification system using Local Binary Patterns (LBP) and Extreme Learning Machine (ELM). The dataset to be classified consists of real marble images collected from the private companies built in organized industrial zone of Elazig. For an effective performance evaluation, we held a comparison with Scale Invariant Feature Transform (SIFT) supported ELM and gave the results. Obtained performance values prove that proposed system is capable for classification of marble textures.