Shape based classification of breast tumors using fractal analysis

In this paper we adopt FFT based fractal analysis method. With the help FFT based method breast tumors (benign and malignant) are classified based on their shapes. In general malignant tumors contour have rough and irregular shapes whereas benign contour have smooth and macrolobulated shapes. In this paper we present a fractal based Fourier transform method to classify the contours. The log -log magnitude frequency response plot and rose plot (for average radial distance and angle variations in terms of average slops and slope of intercept) are used in this study in order to extract the features for classifications.