Medical image classification approach based on texture information

Abstract EN

Texture features play important role in most image classification technique to obtain high accuracy results.

In this work, the medical image classification method considering texture analysis and statistical features have been proposed.

The main concept of proposed method depends on extract statistical features from texture information for each medical image under consideration during classification process.

The proposed classification method consists of two parts.

In first part, the gray level co-occurrence matrix GLCM have been computed for gray medical image and then extract some statistical texture features with second order.

In second part, each of input medical images will be assign to corresponding class depending on GLCM texture features that are extracted in pervious part.

The performance of suggested method evaluated using multi classes for different medical image including heart, liver, and kidney.

The experimental results show ability of proposed method to achieve high accuracy degree in medical image classification.