Automation of snakes in medical images

Abstract

Nowadays time the 3D reconstruction from 2D images has become a tool of daily use for the medical diagnosis. Usually, many tools exist that can make this reconstruction, but when discovering problems in small vessels or areas, as it could be the neck, then tools that detect the borders of the objects of a precise way are needed. For that reasons, in this work, we have used the snake method, that allows to find the contour of small objects with high accuracy. A problem that presents the snake process is that it needs to start from an initial contour, which is usually introduced by hand. This becomes a big problem in medical imaging where the amount of images is very large, and the number of sections of small objects in each image also can be important. In this work we have automated the entrance of this initial contour from an inner point of the object that we want to detect.