Automatic Diagnosis of Breast Tissue
Atef Boujelben, Hedi Tmar, Mohamed Abid, J. Mnif · InTech eBooks · 2012
Will-be-set-by-IN-TECHdepending on edge criterion, solve the problems of segmentation as noise and discontinuities (Osher & Fedkiw, 2002).However, breast quality makes segmentation effective only by taking both intra-area and inter-area aspects into account.To attain our objective which is the ROI segmentation in mammographic images, we apply the Level Set method based on external function (convergence function) that represents area and contour criteria as much as possible.In this paper, we include the texture/shape detection in the process of mammograms diagnosis.The main purpose of this work is the elaboration of a CADi to reach a good identification of ROI and contribute to a better quality of analysis.This work, is integrate within the MIPAX (Medical Image Processing and Analysis eXchange) project which was defined as the object of CES (Computer, Electronic And Smart engineering systems design Laboratory in National School of Engineers of Sfax) and ANIM(Numeric Archiving and Medical Imaging in National School of Medicine) collaboration.So, this project was split in three parts; Numeric Archiving PACS (Picture Archiving and Communication System), Data Base of User Environment and Automatic Analysis of Medical Images.This present work articulate around the last part.To attain our objective (CADi), we firstly show why and how to adapt Level Set-based approach in case of pseudo-detection, which is a semi-automatic detection by using level-set technique; and secondly, we study the performance of boundary, region and texture features in a mammogram diagnosis process.The remainder of this paper is organised as follows.Section 2 presents the state of the art of shape/texture analysis; without loss of generality, we outline the most original and important work addressing mammogram analysis.Section 3 describes the proposed block diagram for mass diagnosis.Section 4 illustrates the deformable model, namely, Level Set approach adopted in segmentation and its adaptation in case of breast cancer detection.Afterwards, section 5 presents the adopted method for analysis and shows how a combination of shape and texture features could be advantageous for a good diagnosis.As for section 6, it presents the results obtained by the proposed scheme.Lastly, section 7 gives some concluding remarks and draws some future work.