期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2006
卷号：6
期号：9A
页码：175-184
出版社：International Journal of Computer Science and Network Security
摘要：This paper presents a scheme of analyzing lung texture to solve the problem of identifying undefined patterns and distinguishing the complex background of superimposed structures in chest radiographs. The method detects and quantifies the interstitial abnormalities in a chest radiograph using the contents based image retrieval (CBIR) scheme. This technique is based on the image feature vector obtained from quasi-Gabor filters and a 3D structure classification scheme. The quasi-Gabor filters are capable of maintaining low computational cost while keeping the important information of the power spectrum of 2D-DFFT, such as band-pass frequency and direction of texture. The 3D classifier is able to capture not only local texture but also global distribution of lung texture and it overcomes the existing problems of a general block operation such as sliding block operation. Our method is a generic one in the sense that it could be applied to analyze the texture of any other images including natural scenery and various medical images such as CT or MRI images
