Rotation and scale invariant three-dimensional recognition using optical scanning holography

We propose a three-dimensional (3-D) image recognition technique using full-color optical scanning holography (FC-OSH) and speed-up robust features (SURF) algorithm. First, we record the complex holograms of objects which are located at different locations. Second, we extract the depth location of the objects from the recorded holograms. Third, we reconstruct the holograms at the object's depth location. Fourth, we match the reconstructed images of the objects with rotation and scale invariance using SURF algorithm. Finally, we extract the 3-D location of the matched object from the matched feature location.