Distortion Analysis in Mammogram Images using Spectral Deviation Mapping

This paper presents a spectral mapping based image coding approach for suspicious region detection using architectural region detection. In the analysis of detecting suspicious regions in mammogram image, segmentation approaches were applied. In the process of region selection, accuracy of decision depends on the coding algorithm, segmentation, and feature descriptor and classification algorithm. The existing approach of suspicious region detection is based on the content, which get distorted during capturing, transmission, or storage process. The distortion added to the image impact the accuracy of region detection and classification. In this paper, to improve the performance of detection, a new spectral domain deviation in textural content of the mammogram region is presented. The direct image mapping of spatial content doesn’t reveal the nature of variation, which tends to misclassification in decision. To improve the accuracy of region selection a spectral domain mapping of energy deviation in textural representation is proposed.