Quantitative Analysis For Three Dimensional Micro-Vessel Imaging

The morphological changes in micro-vessels play an important role in the development of many diseases including cerebrovascular diseases and malignant tumors. With the development of high-resolution imaging, three dimensional micro-vessel imaging has now frequently used in the clinical diagnoses and laboratory research.The accurate extraction of vessel structures and precise quantitative description of their morphology are the essential and most important steps in the application of micro-vessel imaging. However, the variety of vessel structures and the huge amount of data have made it more difficult in processing the images acquired in the three dimensional micro-vessel imaging.A novel image analysis procedure based on micro-vessel imaging including vessel segmentation, thinning and quantification is proposed in this paper.The vessels are segmented from the images using mathematical morphology, and then the skeletons of vessels are extracted using three dimensional thinning. The method of giving topological descriptions is also mentioned. After that, the definition and calculation method of vessel density, branch diameter and tortuosity are given. The algorithm is tested on five experiment data sets including one liver angiographic image, three tumor angiographic images and two brain angiographic images. The result shows that there are distinctive differences in vessel density, tree pattern, diameter and tortuosity among different kinds of vessels.The paper introduces an effective way of analyzing vessel morphology. The algorithm works well in both normal and abnormal vessels of different scales, which proves its great potential for future application in detecting the development of vascular diseases as well as tracking the treatment effect.