Fuzzy image enhancement based microcalsification segmentation from mammogram films

ABSTRACTBreast cancer is one of the most common kinds of cancer in the world. Early diagnose is themost important factor for medical treatment of success for breast cancer which is thefrequently one. Also this rule takes effect to all kinds of cancer. The most useful method ismammogram for diagnosing breast cancer. The role of mammogram is really importantbecause Mammogram?s authentic interpretation means the achievement of the task.Expert?s information and using the linguistic formulation of mathematic give perfectcomment in opportunity with systems of fuzzy. By the help of fuzzy logic we can modelexpressions which change in personal and can not define in classic logic such as ?hot? and?long?.The images of mammogram details which can?t distinguishing with naked eye or analyze canbe discovered by the help of techniques of image processing. But some problems can beoccurred from the changeable of human anatomy in classic methods of image processing.These problems can be solved easily by fuzzy logic. Because fuzzy logic has a flexiblestructure and can give an expert opinion in applications.In this study, an algorithm is developed by using the advantage of image processing methodswhich is changeable grey values, with fundamental rule of fuzzy logic. The aim of developingalgorithm is to reform the image for catching edemas which can not be realized easily inmammogram images. After the application of developing algorithms in the image ofmammograms, these images were tested with seeking the opinion of an expert.Mammogram films those have important functions on diagnosing are interpreted and soresults are by experience and knowledge. Interpreting process on experience may includefaults by human eye or missing on interpreting. Aim of this thesis is generating a system thathelps for this topic and getting better results by comparing their observations and results ofgenerated algorithm.That is important that early diagnosis saves life. And on the other hand the, cost of case thathas early diagnosis is lower than other cases. We realize that film scanning systems aremostly used and density of patients is so high in our country so getting advantages fordiagnosis on short time has been aimed. The aim of this study is generating digital imageprocessing algorithm for facilitating on early diagnosis for breast cancer that increasing in lastyears.Key Words: Breast cancer, mammogram, fuzzy logic, image processing.