Aplikace zpracování biomedicínských obrazů na základě Wavelet transformace

Abstract

The topic of this bachelor thesis is WT (wavelet transform) filtering of different noise from CT images, MRI images, retinal scans and segmentation of filtered image data acquired by CT and MRI. Image noise is an additive component that can obscure important information in an image, which can lead to misdiagnosis. The aim of the work is to test several representatives of several families (types) of wavelets, several levels of image decomposition and subsequent processing and analysing the results. A one quarter of the thesis is devoted to a theory that provides brief information about image filtering. The practical part is divided into two parts. The first is devoted to the processing and analysis of results, which arose from filtering image data by different types of wavelets. The final results of this practical part are wavelet types and decomposition that are best for specific data type, but the worst wavelets are also selected. The second practical part deals with the efficiency of Otsu segmentation after image WT filtering. The goal is to test which wavelet and decomposition level is appropriate to filter a particular type of noise in order to perform the best segmentation. Selected objective evaluation methods are used to evaluate the quality of filtration/segmentation.