Rekha Lakshmanan

Pectoral Muscle Boundary detection - A preprocessing method for early breast cancer detection

Pectoral Muscle (PM), a significant region in Medio-Lateral Oblique (MLO) view of mammogram may adversely affect anomaly detection due to its resemblance to abnormal tissues. The removal of PM region can be considered as a prerequisite step for early breast cancer detection using mammographic images. The principal component of PM boundary component is extracted using the orientation and eccentricity property of Canny edge detected components of coarse mammographic image obtained after a multiscale decomposition technique using Laplacian Pyramid (LP). The principal component of PM boundary is extended to top and left boundaries using nearest neighbor approach. The algorithm was tested on images from the Mammographic Image Analysis Society (MIAS) database as well as mammograms obtained from a representative set of Indian populace provided by Lakeshore Hospital Kochi, India. On comparison with the PM boundary assessed by radiologists, the proposed method yielded an average false positive rate of 0.28%, average false negative rate of 3.67% and low Hausdorff distance for 83 images in mammographic database. Based on the performance analysis of the proposed algorithm, it is observed that 97% of images have an average error less than 3 mm which is promising.

Detection of architectural distortion in mammograms using geometrical properties of thinned edge structures

AbstractThe proposed method detects the most commonly missed breast cancer symptom, Architectural Distortion. The basis of the method lies in the analysis of geometrical properties of abnormal patterns that correspond to Architectural Distortion in mammograms. Pre-processing methods are employed for the elimination of Pectoral Muscle (PM) region from the mammogram and to localize possible centers of Architectural Distortion. Regions that are candidates to contain centroids of Architectural Distortion are identified using a modification of the isotropic SUSAN filter. Edge features are computed in these regions using Phase Congruency, which are thinned using Gradient Magnitude Maximization. From these thinned edges, relevant edge structures are retained based on three geometric properties namely eccentricity to retain near linear structures, perpendicular distance from each such structure to the centroid of the edges and quadrant support membership of these edge structures. Features for classification are g...

A Preprocessing Method for Reducing Search Area for Architectural Distortion in Mammographic Images

Architectural distortion, the third symptom of breast cancer, is considered as the most commonly missed abnormality. The purpose of this work is to detect suspicious regions of interest including spiculation, retraction and distortion in mammograms. Approximately circular homogeneous areas were detected using a high pass isotropic filter. Selected contours were detected from the filtered image to identify suspicious regions. The algorithm was tested on images from the Mammographic Image Analysis Society (MIAS) database as well as 11 mammographic images from a representative set of Indian populace provided by Lakeshore Hospital Kochi, India. The results from the experimental analysis were compared with the information from the radiologist and show the promising scope of the proposed algorithm. The selected regions assist radiologists in analyzing the region with great accuracy.

Enhancement of Microcalcification Features Using Morphology and Contourlet Transform

The proposed method utilizes morphology and contourlet transform for early detection of breast cancer by enhancing micro calcification features. This CAD technique helps the radiologists in reaching a better assessment. The significant edge information indicating the relevant features in various decomposition levels were preserved while removing the artifacts. Target to background contrast ratio, Contrast and Peak Signal to Noise ratio are considered for performance evaluation of the enhancement algorithm. The mini-MIAS mammographic database was employed for testing the accuracy of the proposed method and the results were promising.

Pectoral Muscle Boundary Detection in Mammograms Using Homogeneous Contours

Breast occupies over Pectoral muscle (PM) which is a predominant portion in Medio-Lateral Oblique (MLO) view of mammogram. The similarity in density among PM area and the breast region may generate false positive results which can adversely affect early breast cancer detection. Noise, wedges, opaque markers etc along with labels are unnecessary in mammographic images. The suspicious segments of PM boundary are obtained by extracting contours of homogeneous regions. The geometrical properties of contour segments are analyzed for extracting PM boundary component. An intensity similarity approach extends the detected major PM boundary segment to the two boundaries of mammogram. Experimental analyses were carried out on mammograms obtained from Mammographic Image Analysis database. The proposed methods yields low values for average false positive, average false negative and Hausdorff distance. From the performance analysis of the proposed algorithm, 97% of images have an average error less than 4 mm. Low values of performance measures for the proposed method shows that the extracted PM boundary is close to radiologist drawn PM border.

Straight Line Approximation of Pectoral Muscle in Mammogram Using the Method of Optimum Thresholding

In Medio-Lateral Oblique (MLO) view of mammogram, the presence of pectoral muscle may sometimes affect the detection of architectural distortion due to their similar characteristics with abnormal tissues. As a result pectoral muscle should be handled separately while detecting the breast cancer. The straight line approximation of pectoral muscle is also very important in obtaining the breast tissue architecture, which helps to detect the presence of architectural distortion. In this paper, a novel approach for the straight line approximation of pectoral muscle using optimum thresholding is proposed. The process first selects an optimum threshold based on average gray level and Otsus threshold. The selected region of interest (ROI) of the image is thresholded based on the optimum threshold and approximates the pectoral muscle boundary as a straight line. A set of images was selected for testing and the method is found to have an accuracy of 86.67%.

Investigation of Modulus Maxima of Contourlet Transforms for Enhancing Mammographic Features

Enhancing features of mammographic image assists radiologists in the early detection of breast cancer. In this paper, an enhancement technique using selected modulus maxima of the Contourlet transform is employed to enhance the microcalcification features in mammographic image while simultaneously reducing image noise. Strong edge information corresponding to relevant features was retained based on a parent child relationship among contourlet coefficients at various levels. Simulations were carried out to examine the utility of the proposed technique in mammographic image enhancement. The mini – MIAS database was employed to test the accuracy of proposed method. Contrast improvement index, Peak Signal to Noise Ratio, Target to Background Contrast ratio and Tenengrad Criterion were considered for a evaluating the performance of the proposed methods.

Microcalcification Detection by Morphology, Singularities of Contourlet Transform and Neural Network

-The proposed method presents a new classification approach to microcalcification detection in mammograms using morphology, Contourlet Transform and Artificial Neural Network. Early detection of breast cancer is possible by enhancing microcalcification features obtained using morphology and singularities of Contourlet Transform. The significant edge information indicating the relevant features in various decomposition levels are preserved while removing the artifacts. These features are utilized to detect microcalcifications by classification employing the Back Propagation Neural Network. Target to background contrast ratio, Contrast and Peak Signal to Noise ratio are considered for performance evaluation of the enhancement algorithm. The accuracy of the classification algorithm is 95%. The miniMIAS mammographic database is employed for testing the accuracy of the proposed method and the results are promising. Keywords--Breast Cancer, Back Propagation Neural Network, Contourlet Transform, Morphology