基于乳腺X线图像处理的乳腺癌检测方法研究

项目名称： 基于乳腺X线图像处理的乳腺癌检测方法研究

项目编号： No.61271305

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 无线电电子学、电信技术

项目作者： 陈后金

作者单位： 北京交通大学

项目金额： 80万元

中文摘要： 乳腺癌是女性最常见的一种恶性肿瘤，其发病率正在不断上升，我国己成为乳腺癌发病率增长最快的国家之一。乳腺癌若能在早期得到诊断，可提高患者治疗后的生存期和生活质量。乳腺X线摄影术被认为是最可靠和最有效的乳腺癌检测方法，本课题研究基于乳腺X线的乳腺癌检测方法具有重要意义。 现有检测方法存在诸多问题：主要以美国或英国数据库为研究对象，缺乏针对中国女性常见的致密型乳腺X线图像分析；主要基于单视角乳腺X线图像，缺少多视角的病灶分析；主要进行钙化点和肿块病灶检测，缺少结构扭曲和不对称性等病理特征分析。本课题属于信息学科与医学学科的交叉学科，将针对中国女性高致密性乳腺图像的特点，利用多学科理论来研究乳腺X线图像乳腺癌检测的关键问题，建立基于信息融合思想的乳腺X线多视角分析模型，提出鲁棒的可疑病灶提取与分类方法，有效实现乳腺癌早期筛查与诊断，为临床乳腺癌风险等级评估奠定基础，提高我国乳腺癌检测和诊断水平。

中文关键词： 图像处理；乳腺癌检测；多视角融合；致密型乳腺；风险等级

英文摘要： Breast cancer is one of most common cancers among women and its morbidity is rapidly increasing. Now China has become one of the countries with the highest morbidity of breast cancers. If detected and diagnosed at the early stage, breast cancer is treatable, which will improve the life equality of patients. Mammography is considered as the most reliable and effective method for breast cancer detection. The study on breast cancer detection based on mammography images is greatly significant. There exist some disadvantages in previous detection methods. They were focused more on the American or England mammogram database and less on Chinese women's mammograms mostly with high density, more on single-view mammograms and less on multiple-view mammograms, more on micro-calcifications and masses, less on architecture distortion and bilateral asymmetry. The information discipline and medical discipline will be combined in this project research. According to the characteristics of mammograms with high density, the key problems in breast cancer detection will be studied by use of cross-discipline theory. The multiple-view analysis model will be established based on information fusion. The robust extraction and classification methods for suspicious lesions will be proposed. The screening and the diagnosis will be effective

英文关键词： Image Processing；Breast Cancer Detection；Multiple-view Fusion；Dense Mammogram；Risk Assessment