Variability measurement for breast cancer classification of mammographic masses
Sailesh Gc, Ravi Kasaudhan, Tae K. Heo, Hyung Do Choi · 2015
Breast cancer classification technique divides breast cancer into two categories, benign tumors and malignant tumors. The main purpose of breast cancer classification is to classify abnormalities into benign or malignant classes and thus can help physicians with further analysis by minimizing the possible errors that can be done because of fatigued or inexperienced physician. In this paper, we propose two new shape irregularity measurement based on centroid to contour distance information and measures of variability namely variance and range. Then SVM is used as a machine learning tool to classify breast cancer into two different class (benign class and malignant class) using the shape features extracted using measures of variability. In order to compare the performance of our proposed method with a global shape measure, compactness of each tumor was calculated and compared with the proposed method. The results shows that this new shape feature measurement can improve classification performances for MCC, specificity, sensitivity and accuracy and can aid physicians for undergoing further diagnosis.