Performance Analysis of Segmentation using Thresholding Technique and K-means Clustering Techniques in Mammograms

Abstract

The main objective of the study is the attainment of proper segmentations of mammographic images which are taken from Mammographic Image Analysis Society (MIAS) digital mammogram database. In this study normal breast images and breast image with masses used as the standard input to the proposed system are taken from Mammographic Image Analysis Society (MIAS) digital mammogram database so as to be used in tasks such as to find micro calcification. To full this purpose, we suggest that our work should follow two different techniques, namely, local thresholding technique and K-means clustering technique. Thus, two techniques are used to segment the mammographic images in order to trace the abnormal areas and to find the best one suitable for the objective. Next, we expose in a more detailed manner the objectives proposed in our work to find the micro calcifications in mammograms.