Background: Digital mammography is the most reliable imaging modality for breast\r\ncarcinoma diagnosis and breast micro-calcifications is regarded as one of the most\r\nimportant signs on imaging diagnosis. In this paper, a computer-aided diagnosis\r\n(CAD) system is presented for breast micro-calcifications based on dual-tree complex\r\nwavelet transform (DT-CWT) to facilitate radiologists like double reading.\r\nMethods: Firstly, 25 abnormal ROIs were extracted according to the center and\r\ndiameter of the lesions manually and 25 normal ROIs were selected randomly. Then\r\nmicro-calcifications were segmented by combining space and frequency domain\r\ntechniques. We extracted three texture features based on wavelet (Haar, DB4,\r\nDT-CWT) transform. Totally 14 descriptors were introduced to define the\r\ncharacteristics of the suspicious micro-calcifications. Principal Component Analysis\r\n(PCA) was used to transform these descriptors to a compact and efficient vector\r\nexpression. Support Vector Machine (SVM) classifier was used to classify potential\r\nmicro-calcifications. Finally, we used the receiver operating characteristic (ROC) curve\r\nand free-response operating characteristic (FROC) curve to evaluate the performance\r\nof the CAD system.\r\nResults: The results of SVM classifications based on different wavelets shows DT-CWT\r\nhas a better performance. Compared with other results, DT-CWT method achieved\r\nan accuracy of 96% and 100% for the classification of normal and abnormal ROIs,\r\nand the classification of benign and malignant micro-calcifications respectively. In\r\nFROC analysis, our CAD system for clinical dataset detection achieved a sensitivity of\r\n83.5% at a false positive per image of 1.85.\r\nConclusions: Compared with general wavelets, DT-CWT could describe the features\r\nmore effectively, and our CAD system had a competitive performance.
Loading....