Modified Filtering Scheme for Medical Image

Filtering and feature selection methods for characterization and enhancement of liver and chest images are presented in this paper. Our problem focuses on improvement of performance of classification of feature set by various selection algorithms like sequential forward search (SFS), sequential backward search algorithm (SBS) and sequential forward floating search algorithm (SFFS). Here the focus is on statistical texture features obtained from the computed tomography abdominal images taken by the radiologist. Symlet Wavelet is used for filtering and feature extraction. Wavelet transform decomposes the data in such a way that any hidden information in image can be retrieved. We have tried our best to reduce the feature set size as it is very necessary to limit the number of features or to have optimization of feature set, a design of compact classifier is required for improved classification of the selected problem (Liver Tumor, Chest abnormalities), based on the classification. Most and least significant features are identified on the basis of classification performance of the classifier and then unwanted features may be removed from the set of feature to get a feature subset, which is small in size. In this paper we have tried to characterize the hepatic, hepatoma, hemangeoma, hepatic masses using Symlet Wavelet. After comparing the performance of this technique to that of other approaches it is found to be superior and convenient for medical diagnosis.