基于模糊数的乳腺肿瘤超声图像边缘快速提取方法

乳腺肿瘤边缘的准确提取在临床上对肿瘤良恶性的判别有重要的意义。本文利用三角模糊数的概念，采用重叠式窗口从图像中得到与不同隶属度对应的模糊数，从而建立以步进方格（marching square）为基本单元的模糊数平面；通过区间阈值得到步进方格上的映射区间，根据步进方格算法将对应映射区间着色绘制出肿瘤的边界。分别对恶性和良性肿瘤超声图像进行边缘提取。结果显示，本文方法相比一般提取边缘的算法具有快速准确提取乳腺肿瘤边缘的特点。实验证明本方法可以有效用于乳腺肿瘤超声图像边缘提取。

Fast Edge Extraction for Ultrasound Image of Breast Tumor Based on Fuzzy Number

An accurate edge extraction method for the ultrasound breast tumor image is useful for classifying tumors as benign or malignant. This paper refers to a fast technique to extract edge of breast tumor from ultrasound image. This method uses the triangular fuzzy number to build up a fuzzy number plane whose basic unit is the marching square. It is possible to visualize at once the results obtained using different presumption levels. Experiments of benign and malignant breast tumor in ultrasound images have shown that our method can extract the breast tumor edge faster than many conventional methods can do separately, and the results are reliable and credible. Our experiments demonstrate that it can be efficiently used to extract the edge of breast tumor from the ultrasound image.