Combined Transform domain Based Characterization of Breast Tissues using Nearest Neighbor Classification Techniques

Mammography is a well established imaging technique for showing tissue abnormalities of breast and has been proven to reduce death rate for breast cancer in screened populations of women. The proposed method classifies the breast tissues into normal and abnormal by using Nearest neighbor (NN) classifiers in combined transform domain. The discrete wavelet transform features are extracted from cosine transformed domain. The method is tested on 216 mammogram images from the MIAS database. Experiments on Hybrid transform space shows its potential for accurate mammogram classification. Further, this paper justifies the classification correctness of NN classification techniques which is simple but exploiting the underlying density structure of the dataset. The study reveals the NN variants classification power when combined with cascaded transforms filters feature test.