Kavetha B.V.,eppiaarEngg College |
Chenthil T.R.,eppiaarEngg College |
Chandran G.,eppiaarEngg College
International Journal of Applied Engineering Research | Year: 2015
In this paper, it is aimed to find and detect breast cancer availability in mammogram images. Various researches found to predict breast cancer using various techniques accurately. The existing researches are limited with kind of images, methods, time, automation [partially manual operations] and accuracy in classification. The existing research discussed and makes use of FSand BSmethods for selecting the relevant features and predicting the breast cancer. In this paper, a new approach named ADBM is proposed to find and detect the breast cancer accurately and automatically by utilizing a series of image processing methods. Using Gaussian Filter noise occurred on the image is removed, using Histogram Equalization the contrast level of the image is enhanced and finally using Particle Swarm Optimization method the cancer region is detected from the MIAS database image. Finally the feature values are extracted using GLCM method for classifying the cancer with the help of Multi-SVM classification method and classify the cancer as normal, benign or malignant. ADBM is experimented on MIAS database images using MATLAB software and the results are compared with the SVM-SMO (Support Vector Machine – Sequential Minimal Optimization) technique to show the performance. © Research India Publications.