Feature Extraction Values for Digital Mammograms
Arpana M.A1, Prathiba Kiran2
1Mrs. Arpana M.A, is currently pursuing her M.Tech Degree in Digital Electronics and Communications at AMC College of Engineering, Bangalore which is affiliated to VTU, Belgaum.
2Mrs.Prathibha Kiran, was awarded with M.Tech in Biomedical Signal Processing and Instrumentation during 2011 from Dayanand Sagar College, which is affiliated to VTU, Belgaum.
Manuscript received on May 01, 2014. | Revised Manuscript received on May 03, 2014. | Manuscript published on May 05, 2014. | PP: 183-187 | Volume-4 Issue-2, May 2014. | Retrieval Number: B2250054214/2014©BEIESP
Open Access | Ethics and Policies | Cite
©The Authors. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/

Abstract: Currently digital mammography is the most efficient and widely used technology for early breast cancer detection. The major diagnosing elements such as masses, lesions in the digital mammograms are noisy and of low contrast. The aim of the proposal is to enhance the mammogram images by reducing the noise using median filter, image sharpening and image smoothing. The data clustering algorithm i,e Fuzzy C means clustering is used to segment the region of interest from which various statistical, gradient and geometrical features are extracted. The features extracted from the few images of the data base are used to train the neural networks for classification. The evaluated algorithm is tested on the digital mammograms from the Mammogram Image Analysis Society (MIAS) database. The experimental results show that the breast region extracted by the presented algorithm approximately follows that extracted by an expert radiologist. The detected mass is classified as normal or abnormal. Further abnormal can be classified into benign or malignant.
Keywords: Bio Medical Image processing, Mammograms, Breast Cancer, High Pass Spatial Filter, Fuzzy C means Clustering, Median Filtering, Gradient features.