Segmentation of objects in image data using channel detection

摘要

The present application relates to a computer implemented method of identifying a channel between at least two objects from image data comprising a plurality of pixels or voxels. The method comprises a cycle comprising the steps of choosing a portion of the image data which contains image data of at least a part of at least two of the objects, calculating image gradients of the pixels or voxels of the portion of the image data, and analyzing an orientation of the image gradients. If at least two of the image gradients point into opposing directions, the chosen portion of the image data is determined to contain edges of the objects defining the channel. This method can subsequently be used to separate or segment the objects or parts of the objects defining the channel.