基于灰阶的乳腺肿瘤超声图像特征提取  

Based on Gray-scale Ultrasound Image of Breast Cancer Feature Extraction

作　　者：李莉[1] 吕晓玉[2] 森干[1] 木拉提·哈米提[1] 

机构地区：[1]新疆医科大学医学工程技术学院,新疆乌鲁木齐830011 [2]新疆医科大学第三附属医院,新疆乌鲁木齐830054

出　　处：《医学信息》2011年第18期5973-5974,共2页Journal of Medical Information

基　　金：国家自然科学基金(30960097);新疆医科大学创新基金(XJC201089)

摘　　要：本文根据乳腺超声图像灰度分布的差异,采用灰度直方图对乳腺癌及乳腺增生超声图像进行有效的颜色特征提取,并采用统计学方法进行显著性分析,结果发现乳腺癌和乳腺增生超声图像在峰态和倾斜度上存在显著性差异,具有统计学意义,将有助于进一步对乳腺超声图像进行分类和检索。In this paper, based on differences of breast ultrasound image grayseale distribution, using the histogram of ultrasound images to extract effective color feature of breast cancer and breast hyperplasia, and the use of statistical methods for significance analysis.The results show that ultrasound images of breast cancer and breast hyperplasia in the peak state and tilt, there was significant difference statistically significant, which help to further classify and retrieve ultrasound images.

关 键 词：乳腺超声图像 灰度直方图 特征提取 

分 类 号：R318.04[医药卫生—生物医学工程] TP751.1[医药卫生—基础医学]