Image Enhancement and Its Effects on Segmentation for Mammographic Masses
Breast cancer is a major health problem and continues to be the primary cause of death among women all over the world. Screening mammography is recognized the most effective method for its early detection. Since reading mammograms is an error-prone and time-consume task, a number of computer-aided detection and diagnosis (CAD) systems have been developed to aid the radiologists in the complex work of discriminating types of breast lesions. In almost all of the CAO systems, segmentation of lesions is a very crucial step. Image enhancement which is as a preprocess can largely improve performance of segmentation algorithms. However, the effectiveness of improvements has not been quantized evaluated and compared in previous studies. In this study, we conducted a set of experiments to evaluated two methods, namely 2D segmentation method based on dynamic programming (DPA) and DPA with image enhancement method. The detailed description of our image dataset, experimental procedures and results are presented. The study demonstrates that due to the using of image enhancement, DPA has an obvious improvement in segmenting suspicious regions of interest (ROls) in mammogrphic lesions.
Mammography computer-aided detection and diagnosis image enhancement segmentation performance evaluation dynamic programming
Yong zhang Yihua Lan Haozheng Ren
School of Computer Engineering Huaihai Institute of Technology Lianyungang, China
国际会议
杭州
英文
423-426
2012-10-28（万方平台首次上网日期，不代表论文的发表时间）