Sanjay Nag

Technique for preprocessing of digital mammogram

Digital mammogram has emerged as the most popular screening technique for early detection of breast cancer and other abnormalities in human breast tissue. It provides us opportunities to develop algorithms for computer aided detection (CAD). In this paper we have proposed three distinct steps. The initial step involves contrast enhancement by using the contrast limited adaptive histogram equalization (CLAHE) technique. Then define the rectangle to isolate the pectoral muscle from the region of interest (ROI) and finally suppress the pectoral muscle using our proposed modified seeded region growing (SRG) algorithm. The proposed algorithms were extensively applied on all the 322 mammogram images in MIAS database resulting in complete pectoral muscle suppression in most of the images. Our proposed algorithm is compared with other segmentation methods showing superior results in comparison.

Accurate Breast Contour Detection Algorithms in Digital Mammogram

Aided Diagnosis (CAD) systems have improved diagnosis of abnormalities in mammogram images. The principal feature within the breast region is the breast contour. Extraction of the breast region and delineation of the breast contour allows the search for abnormalities to be limited to the region of the breast without undue influence from the background of the mammogram. After performing an essential pre-processing step to suppress artifacts and accentuate the breast region, the exact breast region as the region of interest (ROI), has to be segmented. In this paper we present a fully automated segmentation and boundary detection method for mammographic images. In this research paper we have proposed a new homogeneity enhancement process namely Binary Homogeneity Enhancement Algorithm (BHEA) for digital mammogram. This is followed by a novel approach for edge detection (EDA) and finally obtaining the breast boundary by using our proposed Breast Border Boundary Enhancement Algorithm. This composite method have been implemented and applied to mini- MIAS, one of the most well-known mammographic database consisting of 322 medio-lateral oblique (MLO) view obtained via a digitization procedure. To demonstrate the capability of our segmentation algorithm it was extensively tested on mammograms using ground truth images and quantitative metrics to evaluate its performance characteristics. The experimental results indicate that the breast boundary regions were extracted accurately characterize the corresponding ground truth images. The algorithm is fully autonomous, and is able to preserve, skin and nipple (if in profile), a task very few existing mammogram segmentation algorithms can claim.

MAMMOGRAPHIC DENSITY ESTIMATION AND CLASSIFICATION USING SEGMENTATION AND PROGRESSIVE ELIMINATION METHOD

For establishing risk factor of breast cancer requires highly specific breast density measure that can result in a more focused breast cancer prevention, diagnosis and treatment. This paper proposes a new CAD system for density estimation using progressive elimination method. The lower intensity pixels are eliminated in multiple phases by targeting specific intensity bands in each phase, using established statistical techniques. Local Standard Deviation (LSD) values are used to identify significant transitions and MLSD values to isolate the most significant transitions or edges. The results are compared to ACR BI RAD system of classification to establish the risk factor. Accuracy estimation on the proposed segmentation method signifies satisfactory qualitative results. The proposed algorithm implemented on all 322 mammograms of MIAS shows 73.91% agreement. The obtained Kappa (κ) value for the proposed method is 0.673.

A Novel Approach to Detect Accurate Breast Boundary in Digital Mammogram Using Binary Homogeinity Enhancement Algorithm

Computer Aided Diagnosis (CAD) systems have improved diagnosis of abnormalities in mammogram images. The principal feature within the breast region is the breast contour. Extraction of the breast region and delineation of the breast contour allows the search for abnormalities to be limited to the region of the breast without undue influence from the background of the mammogram. After performing an essential pre-processing step to suppress artifacts and accentuate the breast region, the exact breast region as the region of interest (ROI), has to be segmented. In this paper we present a fully automated segmentation and boundary detection method for mammographic images. In this research paper we have proposed a new homogeneity enhancement process namely Binary Homogeneity Enhancement Algorithm (BHEA) for digital mammogram. This is followed by a novel approach for edge detection (EDA) and finally obtaining the breast boundary by using our proposed Breast Border Boundary Enhancement Algorithm. This composite method have been implemented and applied to mini-MIAS, one of the most well-known mammographic databases consisting of 322 mediolateral oblique (MLO) view obtained via a digitization procedure. To demonstrate the capability of our segmentation algorithm it was extensively tested on mammograms using ground truth images and quantitative metrics to evaluate its performance characteristics. The experimental results indicate that the breast boundary regions were extracted accurately characterize the corresponding ground truth images. The algorithm is fully autonomous, and is able to preserve skin and nipple (if in profile), a task very few existing mammogram segmentation algorithms can claim.