Tumor diagnosis and screening plays vital role in treating the patient from life threatening illness. The Mammography is the early diagnosis and screening method of taking low energy level x-ray images of the human breast to detect any masses or microcalcifications present which are the early sign of cancerous tumors. Expert Medical Practitioner is required in classifying these tumors as cancerous or non-cancerous. In this paper we proposed a tumor classification and extraction method from Mammogram images using Convolutional Neural Network (CNN).With the help of this proposed method we are extracting and classifying the tumor as malignant (cancerous) or benign (cancerous). We developed a Computer Aided Diagnosis (CAD) system with CNN for human breast tumor extraction and classification. The Proposed method is performed in different phases namely Preprocessing by an Adaptive filter, Segmentation by the Gaussian mixture model (GMM), Feature extraction by the relationship of the pixels in Spatial Domain with the help of Gray-Level Co-Occurrence Matrix (GLCM). and finally classification by CNN Classifier. The proposed method experimental results shows greater accuracy of 98.46% in classifying human breast tumor as malignant or benign compared to SVM Classifier. Also the interactive visual system by Graphical User Interface (GUI) designed in MATLAB facilitates the proposed system to be operated on several Mammogram images.
Volume 12 | 07-Special Issue
Pages: 2636-2641
DOI: 10.5373/JARDCS/V12SP7/20202400