chapter
6 Pages

Texture feature extraction and classification for CT image of xinjiang local liver hydatid based on support vector machine

WithF. Yang, M. Hamit, C.B. Yan, A. Kutluk, W.K. Yuan, E. Alip & A. Matmusa

Hepatic hydatid disease, which often occurs in the pastoral area, is an infection of larval stage animal tapeworm. Xinjiang Uygur Autonomous Region, a multi-ethnic province in northwestern of China, is one of the most important foci of hepatic hydatid disease in the world [1]. CT examination is an effective method for early diagnosis and disease screening [2]. Texture, which is a basic property of the object surface, is widely spread in the nature world. It has important significance for the description and recognition of the object [3]. Texture feature is independent of the color and brightness to reflect the homogeneous phenomenon of visual characteristics [4]. It has been widely used in pattern recognition and computer vision. Fang DONG [5] proposed the texture feature extraction of liver CT images based on fractal dimension, which pointed that the fractal dimension value of the soft tissue in the same organs was just related to the property of organs. The fractal dimension value is different between the normal liver and

liver cancer, which can separate the normal liver from the cancer tissue. Stavroula G. M. defined an optimal performing computer-aided diagnosis architecture for the classification of liver tissue from non-enhanced CT images, five distinct sets of texture features were extracted using first order statistics, spatial gray level dependence matrix, gray level difference method, Laws’ texture energy measures and fractal dimension measurements [6]. Doaa Mahmoud-Ghoneim investigated the accuracy of texture analysis results on three color spaces, conventional grey scale, RGB, and Hue-SaturationIntensity (HSI), at different resolutions [7]. Zhang Gang extracted the texture features based on Gabor wavelet and investigated the classification of liver disease [8].