Abstract: Blood vessels segmentation and visualization is an important task since it is used to analyze the health status of a biological tissue. However, this task becomes difficult because of the noise in the image, mainly at high depths. This paper addresses a method based on a morphological analysis of blood vessels for segmentation in LSI images. Moreover, the proposed methodology deals with the disconnection in the blood vessel region, caused by intense noise, restoring the structure to allow its visualization. Results demonstrate that it is possible to visualize blood vessels even at 600μm, and that the proposed reconnection map also is useful for contrast improvement.
0 Replies
Loading