Result
As we can observe from the stitched images the orb descriptors are invariant to viewpoint changes. It is also faster to compute the Orb matching than the floating point based HOG like descriptors
Try it for yourself with the
CamApp
Back to index