Comparative Study of Wavelet De-noising Threshold Filters for Mammogram Images Classification Based on Fuzzy Soft Set Theory

Noise present in the digital mammograms directly influences the capability and competence of a classification task which makes de-noising a challenging problem. In the literature, few wavelets like daubechies db3 and haar have been used for de-noising medical images. Nevertheless, wavelet filters such as sym8, dB3, dB4, haar and Coif1 at certain level of soft and hard threshold functions have not been taken into account for mammogram images. Therefore, in this study two wavelet filters namely: sym8 and daubechies db3 at certain level of soft and hard threshold functions have been considered for classification of mammogram images. Meanwhile, in terms of mammogram images classification using data mining methods review on literature showed that no work has been done using fuzzy soft set based similarity measure for classification of mammogram images. Therefore, the positive reviews produced from past works on fuzzy soft set based classification have resulted in an attempt to use fuzzy soft set for mammogram images classification. Thus, the proposed methodology involved five steps namely MIAS dataset, images de-noising using wavelet hard and soft thresholding, region of interest identification feature extraction and classification. Hundred and twelve images (68 benign images and 51 malignant images) were used for experimental set ups. Experimental results show better classification accuracy in the presence/absence of noise in mammogram images where the highest classification rate occurs with db3 (level 4) with accuracy 62.12 % (soft threshold) with CPU time 0.0026sec and classification rate without noise 63.64%.