Fractal Features based on Differential Box Counting Method for the Categorization of Digital Mammograms

Abstract

Computer aided diagnostic systems can assist radiologist in detecting breast cancer at an early stage with improved mammogram interpretation efficiency. In this paper, six fractal based features obtained from the fractal dimension computed using differential box counting method, are used for distinguishing between normal mammograms from the cancerous ones. The new fractal feature f6 derived from the modified average image is found to be a better feature for distinguishing between normal, malignant and benign masses and mammograms with microcalcifications. The average values of the new normalized fractal feature for normal, mammogram with microcalcifications, benign and malignant tumors are obtained as 0.125, 0.4737, 0.2954, and 0.5992 respectively. The area under the Receiver Operating Characteristics (ROC) curve is found to be 0.923. The study is validated using the mammograms obtained from the online Mammographic Image Analysis Society (MIAS) Digital Mammogram database.