Surface reconstruction for efficient colon unfolding

Unfolding is a new visualization method for colorectal disease and polyp detection. Compared with a virtual endoscopy method, it is more suitable for medical applications because it gives us unfold images of the inner surface in an organ. However, since conventional unfold methods generate only 2D images, it is difficult to show the surface at a first glance and to manage unfolding results with diverse viewing controls such as rotation and magnification. To solve it, we propose an efficient unfolding method using surface reconstruction. Firstly, we generate a 2D unfold image using volume ray casting. At the same time, we store distance values from each sample point on a central path to colon surface, for all rays. After making a height field from the distance values, we reconstruct a 3D surface model. Lastly, a 3D unfold image is acquired by mapping the 2D unfold image into the 3D surface model. Since our method offers the overall shape of an organ surface, the problematic areas can be identified quickly and inspected afterwards in more detail.