Digital Mammogram Segmentation using Non - Shannon Measures of Entropy

Abstract— Mammogram analysis usually refers to processing of mammograms with the goal of finding abnormality presented in the mammogram. Mammogram segmentation is one of the most critical tasks in automatic mammogram image analysis. Main purpose of mammogram segmentation is to segment suspicious regions by means of an adaptive threshold. In image processing, one of the most efficient techniques for image segmentation is entropy-based thresholding. This approach uses the Shannon entropy originated from the information theory considering the gray level image histogram as a probability distribution. In this paper, non-Shannon measures (Havrda & Charvat, Renyi and Kapur) of entropy are applied to segment mammograms and their comparison has been done with Shannon entropy. Main advantage of nonShannon measures of entropy over Shannon entropy is that non-Shannon measures of entropy have parameters ( in case of Havrda & Charvat and Renyi and ,  in case of Kapur) that can be used as adjustable values. These parameters can pay an important role as tuning parameters in the image processing chain for the same class of images. Experiments have been conducted on images of mini-MIAS database (Mammogram Image Analysis Society database (UK)). In order to validate the results, several different kinds of standard test images (fatty, fatty-glandular and denseglandular) of mini-MIAS database are considered. Some typical results are presented to illustrate the influence of the parameters in the thresholding. It has been observed that nonShannon measures of entropy work effectively for different kind of mammograms. The Results of this study are quite promising. This study can be a part of developing a computer aided decision (CAD) system for early detection of breast cancer.