Computer-Aided Diagnosis Scheme Using a Filter Bank for Detection of Microcalcification Clusters in Mammograms

Abstract

Mammography is considered the most effective method for early detection of breast cancers. However, it is difficult for radiologists to detect microcalcification clusters. Therefore, we have developed a computerized scheme for detecting early-stage microcalcification clusters in mammograms. We first developed a novel filter bank based on the concept of the Hessian matrix for classifying nodular structures and linear structures. The mammogram images were decomposed into several subimages for second difference at scales from 1 to 4 by this filter bank. The subimages for the nodular component (NC) and the subimages for the nodular and linear component (NLC) were then obtained from analysis of the Hessian matrix. Many regions of interest (ROIs) were selected from the mammogram image. In each ROI, eight features were determined from the subimages for NC at scales from 1 to 4 and the subimages for NLC at scales from 1 to 4. The Bayes discriminant function was employed for distinguishing among abnormal ROIs with a microcalcification cluster and two different types of normal ROIs without a microcalcification cluster. We evaluated the detection performance by using 600 mammograms. Our computerized scheme was shown to have the potential to detect microcalcification clusters with a clinically acceptable sensitivity and low false positives.

A New CAD System for Breast Microcalcifications Diagnosis

“…There are other techniques that have been exploited in breast lesion characterization such as Gabor filter bank [23,57], Gaussian Laplacian filter [59] and fractal dimension [60].…”

“…Several classifiers were used to distinguish malignant lesions from benign ones [14,60], but the most commonly used are Neural networks [23,35,39], K nearest neighbors [44], Bayesian classifier [39,57,61], Quadratic classifier [52], Linear classifier [20], Expert system [18], Binary decision tree [71], Genetic algorithms [11], SVM [25,58] and Adaptive thresholding [67].…”

A new algorithm for automatic vascular mapping of DCE-MRI of the breast: Clinical application of a potential new biomarker

“…In order to enhance and extract vessels on each smoothed image, the morphology of curvilinear structures was assessed via eigenvalue analysis of the Hessian matrix. After setting all positive eigenvalues to zero, the smallest eigenvalue was subtracted from the largest eigenvalue, allowing to distinguish linear structures, which have positive values, from nodular structures, having values that approximate zero [13,18]. For a fair comparison of images at multiple scales, for each scale the obtained subimage was multiplied by the respective value (normalization).…”

Wavelet Transformation for Enhancing Mammographic Images

“…These methods are usually composed of a wide range of combination of fuzzy logic, wavelet transformation or that of the neural network [2][3][4][5][6]. Since the mammograms show larger areas of varying contrast and brightness, thus the information is highly susceptible to being correlated [7][8][9][10]. Other researchers used wavelet transformation to an extent where it tends to give more consolidated results than the other methods [11,12]; therefore, the following study presents an effectively modeled algorithm for multi-wavelet transformation to denoise the noisy mammographic images to allow easy microclassification to help doctors or radiologist detect breast cancer easily.…”