Automatic identification breast cancer using multiresolution algorithm

Abstract

In this paper, we present a multiresolution scheme to detect stellate lesions in mammograms. Multiresolution analysis is used to analyze the images at different frequencies with different resolutions. First we removed the noise from mammograms using Multiresolution analysis and then we detected tumors. Then, using the Embedded Zerotree Wavelet (EZW) algorithm, we compressed denoising mammographic image and showed that by applying the algorithm to detect tumors in the compressed image we obtained the same results as in the case of non-compressed images. Experimental results obtained from the mammographic images of patients recorded in the Clinical Center in Kragujevac, show that using multiresolution algorithm can be detected tumors of different sizes.