Impact of undistortion and interpolation on fiber optic tracking

Fiber optics are commonly used for image transmission in image guides of flexible endoscopes which are utilized in various medical and industrial tasks. Many existing algorithms for video coding like motion compensation, Super-Resolution or object tracking could improve the usability and quality of these devices. Since the crucial aspect of these algorithms is detecting and tracking features between successive images we set up a framework to automatic parametrize and evaluate motion tracking for various environments and textures. We therefore adapted a state of the art tracking approach to suitably work for the field of fiberoptics and can quantitavely confirm the positive effect of interpolation and undistortion of acquired image data. Results for four kinds of image content are grafically given in meaningful plots as well as summarized in a detailed table.