Mammographic Image Analysis of Breast Using Neural Network
Lesa MAMBWE, Chitundu BWALYA, Sudha RAMASAMY · DOAJ (DOAJ: Directory of Open Access Journals) · 2015
This paper discusses the various stages of detecting tumours of the breast mammogram images. A Neural Network algorithm is applied for obtaining the complete classification of the tumour into normal or abnormal. The most important procedure or technique for obtaining the classification is the feature extraction, by extracting a few of discriminative features, first-order statistical intensities and gradients. The Image Pre-processing technique is essential prior to Image Segmentation in order to obtain accurate segmentation. Thus mass detection can be carried out. The processes involved in achieving the three techniques mentioned above include global equalization transformation, denoising, binarization, breast orientation determination and the pectoral muscle suppression. The presented feature difference matrices could be created by five features extracted from a suspicious region of interest (ROI). Grey Level Co-occurrence Matrix (GLCM) aids the obtaining of statistical features such as correlation, energy, entropy and homogeneity. The other statistical to features to obtain are area, moment, variance, entropy, standard deviation and moment. The Neural network technique yields results of abnormal mammograms.