Tracking of the breast cancer by Microcalcifications segmentation based on Wavelet transform AHMED REKIK, MOHAMED SALIM BOUHLEL National Engineering School of Sfax Laboratory of Electronics and Information’s Technologies (LETI) B.P.W, 3038 Sfax, Tunisia Abstract: It has long been hoped that the power of computers could be harnessed to detect early subtle signs of breast cancer, including calcifications and speculated masses on mammograms. This was achieved with the development of commercially available CAD systems that search the complex background of dense breast tissue and fat for bright calcifications or lines in speculated masses. In this context, a large variety of segmentation algorithms have been developed. This paper presents an objective study of segmentation algorithms using the wavelet transform for the detection of microcalcifications. The results of implementation and an evaluation of different wavelet function are also presented. Key-Words: breast cancer, digital mammography, wavelet, segmentation, Multiresolution approach, filter banks approach, microcalcifications. 1 Introduction: Recent studies show that breast cancer is the leading cause of death of women between the age of 35 and 54 [1]. Early detection is the most successful method of dealing with this epidemic. Currently, the best method for early detection is the use of mammography. Other techniques, such as computed tomography, magnetic resonance imaging, and ultrasound, have been investigated, but mammography remains the proven technique [2]. Indeed, the best means of struggle against this illness remains the prevention by the precocious tracking, and this through the intermediary of mammographic exams. The aim of mammographic exam is the research of possible suspected radiological signs, among them, microcalcifications which correspond to a small anomalous limestone deposits. Microcalcifications are distinguished from the normal structures of the mammary gland by a certain number of variable features according to the topic and the nature of microcalcifications: the shape, the contrast, the contour, the texture.... In this context, the implementation problem of an automatic methods detection system resides globally in the difficulty to specify the problem and to automate the segmentation and the detection of these signs. The automatic analysis of mammographic images is a problem that has not been landed that enough lately, since the first works in this domain are less than ten years old [3]. Some previous publications proved the interest of an automatic analysis system in the setting of the microcalcifications detection [4] [5]. In this context, our work consists to the exploitation of wavelet transform for the development of microcalcifications detection algorithms. An assessment of results brought by the different wavelets types based on these two approaches, will put in value the interest of this work. 2 Description of our approach: 2.1 Position of the problem Mammographic images reveal a relative contrast between two principals constituent of the breast: the greasy cloths and the conjonctivo-fibrous matrix. In general manner, it’s extremely delicate to define a normality of mammographic images: indeed, the aspect of the mammary gland is