Texture Characteristics for Classification of the Ultrasonic Images of Roator Cuff Disease

Abstract

This article proposes a study on applied the texture analysis method to classify the different disease groups that are normal, tendon inflammation, calcific tendonitis and rotator cuff tear. The supraspinatus tendon is usually involved among above-mentioned diseases progression. Four texture analysis methods that texture feature coding method, gray-level cooccurrence matrix, fractal dimension and texture spectrum are used to extract features of tissue characteristic of supraspinatus tendon. The mutual information method is independently used to select powerful feature among four texture analysis method, further, the radial basis function network to classify the ones into the four disease group. Experimental results tested on 85 images reveal that the proposed system can achieves 84% accurate rate.