3D Construction of Endoscopic Images Based on Computational Stereo

Endoscopic surgery, a typical example ofminimally invasive surgery, has changed the traditional surgicalprocedure in the way of accessing the visual informationavailable on the corporal site. Its development heavily relies onthe improvement of the visualization technologies. Three-dimensional (3D) image system that is capable ofproviding more depth cues, therefore, has a potentially strongimpact on the current 2D image guiding technology. Unlikemost other systems that utilize stereoendoscopes with dual lens, we study the feasibility of 3D image construction fromsingle-lens endoscopic video by exploring the theory of multipleview geometry. We coupled a regular endoscope to aposition-measurement equipment to provide extrinsicparameters of the acquired images. With the available multipleview geometric algorithms, depth maps were constructed fromtwo consecutively collected images. We present theexperimental results and conclude that computational stereopsishas a great potential in building 3D endoscopic image systemsfor both surgical aiding and training.