Texture Analysis of B-Scan Image Using Fractal Dimension
Liu Fang Cheng Jingzhi
(Department of Biomedical Engineering and Instrumentation,Xi'an Jiaotong University,Xi'an 710049)
Abstract Based on the texture analysis of B-Scan image fractal model, this paper presents the liver clinical B-Scan images and their fractal characteristics. It describes the semi-fractal characteristics of B-Scan images and calculates the fractal dimension(D) of the liver B-scan images of 10 normal persons,3 patients with cancer of the liver, and 3 patients with cirrhosis of the liver. The results indicate that Fractal Dimension (D) is an important parameter in differentiating some diseases. On the basis of the influence of Fractal Dimension (D) and scale (ε),this paper suggests the texture classified parameter S. The experiment results indicate that the parameter S can be used to differentiate diseases efficiently.
Key words Texture analysis Ultrasonic image Fractals Image processing
1 引 言
超声诊断技术，因其具有无痛、无损、无离子幅射和可重复检查等优点，越来越广泛的应用于临床诊断。提高人体组织超声图像识别的可靠性、科学性，是发展这一技术应用的重要研究课题。近年来国内外不少学者做过这方面的研究，取得了一定的进展。有代表性的纹理分析方法是：一阶统计方法［1］，二阶统计方法与自相关函数法［2，3］，空间灰度层共现矩阵法［4］，随机区域模型法［5］。但是，由于课题本身的复杂性，目前还没有有效的方法应用于临床，尚待进一步研究。
[1] 2 3 4 5 6 7 下一页
