Breast Cancer Classification Using Threshold Segmentation and Voting Classification Method

The complex issue before professionals is to compute the mammography in these days, and Computer Aided Diagnosis (CAD) systems are required for this task. The mammography becomes a robust tool for diagnosing the breast cancer. Subsequent to attain the mammograms, the major task is to analyze the images for determining the presence of cancer. However, detecting breast cancer is quite challenging task in machine learning and image processing. Moreover, the morphology of the calcifications plays a significant role in classifying the breast images as benign or cancerous. Various stages are executed for detecting breast cancer. This work enforces Otsu’s segmentation along GLCM algorithm to extract the features. The earlier work has implemented Random Forest (RF) to classify the cancer. This work suggests a Voting classification algorithm to classify the breast cancer accurately. In this algorithm, logistic regression, naive bayes and random forest models are integrated. According to the simulation results, the suggested algorithm has better accuracy, precision and recall than existing methods. The proposed model is compared with the random forest for classifying the breast cancer. The proposed model shows the significant improvements in the results in terms of accuracy, precision and recall.