Image Texture Analysis based on Color Level Methods using Neural Network

Texture and colour are very important elements in image analysis. This article provides advanced research on the rapid acquisition and classification of coloured textures. In our research we use the principle of Grey Level Co-occurrence Matrix (GLCM) and its modifications. The main idea was to propose a method to describe the combined elements, precisely colours and textures. This work consists of our previous research on the Colour-level co-occurrence matrices (CLCM). The CLCM method is modified by a one-dimensional GLCM approach. The texture classification process is performed using a robust vector machine auxiliary classifier. We also conducted several experiments with convolutional neural networks (CNN) with 15 layers and AlexNet with 25 layers. Finally, all evaluations and comparisons of the colour texture acquisition results for all methods used are given. All algorithms were tested on two different known colour texture databases (Outex and Vistex databases).