Image Sequence Analysis for on line Observation of Thrombus Formation in Blood Vessels.pdfVIP

Image Sequence Analysis for on line Observation of Thrombus Formation in Blood Vessels Ralf TÖNJES Institut für Theoretische Nachrichtentechnik und Informationsverarbeitung, Division: Automatic Image Interpretation (Prof. C.–E. Liedtke) Universität Hannover, Appelstr. 9A, 30167 Hannover, Germany, E–Mail: toenjes@tnt.uni–hannover.de Abstract. An image analysis system for in vivo observation of thrombus formation in skin blood vessels of small marmals is presented. A model based approach is applied, allowing automatic segmentation of thrombus, blood vessel, and tissue in low contrast images. Image sequence analysis takes advantage of preceeding results in three ways: data reduction, improvement of algorithm efficiency, and robustness. This is achieved by a fast motion compensation that takes the image movements in the in vivo observation into consideration. For clinical use in a biomedical laboratory a graphical user interface was developed. Key words: biomedical image processing, image sequence analysis, model based image analysis, motion com- pensation, real time analysis 1. Introduction 2. The Model In thrombus research the in vivo observation of blood Two types of a priori knowledge can be distinguished, vessels in the skin of small marmals is used to study thrombus describing static and dynamic features concerning thrombus formation. After triggering thrombus formation by a con- formation. trolled local injury at the inner blood vessel wall the genera- For segmentation (chapter 3) the static features F1 to F3 tion and temporal development of the thrombus can be ob- are relevant: served under the microsco