Clustered calcification analysis and detection for mammographic images based on statistical texture models

In this paper, an algorithm for texture analysis of clustered calcification based on statistical texture models is proposed. The prior knowledge of both normal and lesion training samples are incorporated into statistical texture models separately. Specifically, beside texture analysis of the lesion tissues, and the resultant statistical parameters can also be used for unknown sample representation and classification. The experimental results show that the proposed method has better performance than the traditional SVM based classifiers. The proposed method can also be applied into other types of medical image analysis and classification.