Dealing with Abnormalities in Mammogram by Using Wavelet Analysis

The research aims to create a technique for categorizing regions of interest (ROI) in mammograms that encompass a variety of breast tissue types, including both normal and abnormal. In this study, we have used bi-orthogonal transform for the classification. The first part of study includes image preprocessing of mammograms. The next half includes convolution of various transforms with ROI, feature extraction and classification. To find better classification results, the scheme is tested for various transforms. The support vector machine (SVM) is used to validate the presence of abnormalities in mammograms. The proposed system resulted to give maximum accuracy 95.85% for RBF kernel and sensitivity 97.28% at linear kernel. The overall study recommends that the bi-orthogonal transform and SVM can be useful for identification of abnormality in mammogram. This study will aid researchers in discovering more effective high-performance methods, benefiting radiologists as well. The system's contribution will facilitate societies in pursuing improved treatment options for women globally, potentially reducing the mortality rate.