Unified Preprocessing and Enhancement Technique for Mammogram Images
Sushreeta Tripathy, Tripti Swarnkar
Abstract
In developed countries, breast cancer is one of the foremost reasons for the increase in mortality among women. Microcalcifications in breast tissue are one of the key indications appraised by the radiologist for identification of breast cancer in its early stage. To identifying such microcalcification, masses and architectural distortion in breast preprocessing in mammogram plays a vital role. Additional imaging provides bit more information than an initial screening, and more focus is made to the sceptical masses. For this cause, preprocessing of mammogram images is essential in the process of breast cancer examination as it could reduce the rate of false positive. Prior diagnosis of cancer and other abnormalities in human breast, digital mammogram has appeared as the most accepted screening approach. This manuscript presents contrast enhancement, by using the contrast limited adaptive histogram equalization (CLAHE) and thresholding methods for detecting the breast tumor boundaries from digital mammogram. The proposed techniques were applied in the MIAS database, which contains 322 mammogram images. The breast enhancement and segmentation technique by using thresholding provides promising results. To compare the performance of the studies, contrast improvement index (CII) is used as a performance evaluation parameter.