An Investigation of Bayes Algorithm and Neural Networks for Identifying the Breast Cancer

• Published in: Indian journal of medical and paediatric oncology Vol. 38; no. 3; pp. 340 - 344
• Main Authors: Udayakumar, E, Santhi, S, Vetrivelan, P
• Format: Journal Article
• Language: English
• Published: A-12, 2nd Floor, Sector 2, Noida-201301 UP, India Thieme Medical and Scientific Publishers Pvt. Ltd 01.07.2017; Medknow Publications and Media Pvt. Ltd; Medknow Publications & Media Pvt. Ltd; Medknow Publications & Media Pvt Ltd
• Subjects: Accuracy; Algorithms; Artificial intelligence; Artificial neural network; Breast cancer; Cancer; Classification; computer-aided diagnosis; Diagnosis; gray-level co-occurrence matrix; Investigations; Lung cancer; mammogram; Mammography; Medical imaging equipment; Motion pictures; Neural networks; Original; Original Article; Parameter estimation; Pattern recognition; region of interest; Womens health; Artificial neural network; mammogram; gray-level co-occurrence matrix; computer-aided diagnosis; region of interest
• ISSN: 0971-5851; 0975-2129; 0975-2129
• DOI: 10.4103/ijmpo.ijmpo_127_17

Abstract Context: Breast cancer is a biggest threat to women. X-ray mammography is the most effective method for early detection and screening of breast cancer. It is a tough challenge for the radiologist in reading mammography since it does not provide consistent result every time. Aim: To improve the primary sign of this disease, computer-aided diagnosis schemes have been developed. Using monitor, digital images of mammography are displayed and they can be lightened or darkened before they are printed on the film. Time factor is important to identify the abnormality in body such as breast cancer and lung cancer. Hence, to detect the tissues and treatment stages, image-processing techniques are improved in several medical areas. In this project, using low-level preprocessing techniques and image segmentation, the breast cancer detection is done. Methods: With the help of Bayes algorithm and neural networks (NNs), the type of the mammogram and stages is identified. For segmentation process, region-growing algorithm is used, which helps to find the affected portion, i.e., region of interest. Gray-level co-occurrence matrix (GLCM) and texture feature are used for feature extraction. Results: Bayes algorithm is used for probability of identification, whereas NNs is used to reduce the probability level from 0–1000 to 0–1 in case of classification.