Fractal Analysis of Microscopic Images of Breast Tissue

This paper presents a new method for the detection of abnormalities in microscopic images of breast tissue. Breast cancer is the most common form of cancer among women. The diagnosis is confirmed by the subjective evaluation of tissue samples, that is usually done by the pathologist manually. An automatic system for the analysis of tissue samples would be of great help. The aim of this study was to investigate the feasibility of a new method for characterizing tissue images. The method is based on fractal geometry. Fractals are geometrical entities which are well suited to capture the complexity and the variability of biological structures. We estimated the fractal dimension using the euclidean distance mapping. Reported results show that the value of fractal dimension is a very good feature in separating normal from cancerous tissue images. Key-Words: Medical images, breast cancer, fractal analysis, quantitative pathology, classification