Analýza medicínských snímků za účelem rekonstrukce objektu na snímku k následnému zpracování

Abstract

This study deals with the methods of analysis and processing of medical images and their use at the reconstruction of the carotide artery at the ultrasound images in B-mode, transversal section. The aim of this study si to project the methos for autoatic and semiautomatic detection of the inner artery wall and o implement these procedures subsequently in the Java language to the fotogrammetric system FOTOMNG , which is built on the NetBeans modular platform. The attention is paid mainly on the filtering of the ultrasound images and the active contours methods, both parametric and geometric. Designed algorithm then uses the Hough transform and gradient vector flow (GVF) as an external forces that affect the deformable models. Thanks to the tool integration to the FOTOMNG system, this is possible to make the the quantitative analysis of the detected objects immediately, thus to be able of facilitate markedly the work of the physician.