Optical-Flow Perspective Invariant Registration

In this paper we examine the causes of one of the major shortcomings of current natural feature registration approaches, failure to register when the camera's view approaches parallel to the marker. The methods used by current registration algorithms in the attempt to overcome this problem are reviewed, and a novel tracking based approach called the Optical-flow Perspective Invariant Registration Augmentation (OPIRA) is presented which significantly improves the range of registration. A thorough evaluation of OPIRA is conducted using an external ground truth, showing the improvements possible when combined with leading natural feature registration algorithms such as SIFT, SURF and Ferns.