Analysis of breast tomosynthesis images

OData support
Dr. Horváth Gábor
Department of Measurement and Information Systems

The main goal of the thesis would have been presenting the possibility of the Digital Breast Tomosynthesis (Tomo) CAD systems and the detection of a well-defined family of pathological lesion. The Tomo is a three-dimensional imaging technology that makes a few dozen x-ray image in a limited range of angle about the breast. After that owing to the computer synthesis we get up to 50 parallel reconstruted slices of the breast volume. They are more detiled than the projected views of the conventional mammography and these slices provide three-dimensional information about the breast. The target lessions are the small microclacification clusters. The lack of Tomo database and on account of other matter, that will be explained in the thesis, during the semester conventional MLO and CC X-ray images were used. I divided the task into two main section, the first is a clacification detection task and the second is a fals reduction part. But only the first part had been implemented even not completely. It consist of three stage, first an masking method, whish makes one-dimensional curves of the two-dimensional image and amplifie some entity of it, the second stage is a wavelet transform and the last part is the filtering of result of stage 2.


Please sign in to download the files of this thesis.