Classification of Clustered Microcalcifications in Mammograms Using Graph Method
semanticscholar(2016)
摘要
Novel method for the classification of micro calcification clusters in mammograms is proposed Methods. The topology/connectivity of individual Morphology of micro calcification clusters, such as cluster area, cluster perimeter, cluster diameter, cluster circularity, cluster eccentricity, and cluster elongation. The radius of the structuring element is equal to 6 pixels (i.e., scale=6). The boundaries of dilated micro calcifications are displayed using different colors and each individual micro calcification is labeled with a sequential number which is ordered according to the spatial location of the corresponding micro calcification in the image patch. A set of micro calcification graphs were constructed to describe the topological structure of micro calcification clusters at multiple scales. In a micro calcification graph, each node represents an individual micro calcification, and an edge between two nodes is created if the two corresponding micro calcifications are connected or overlap in the 2-D image plane. Artificial Neural Network (ANN)-based classifiers are used for classifying micro calcification clusters into malignant and benign. Index Terms – MCCMicro calcification, ANNArtificial Neural Network, GUIGraphical User Interface.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要