Performance & Comparison of Surf and SIFT using RSOC in Image Registration

This paper present performance analysis comparison of Speeded-Up Robust Features (SURF) and scale invariant feature transform (SIFT) using Restricted Spatial Order Constrains (RSOC) in image registration. The SURF differ from SIFT for using integral images i.e. some of all pixel for image convolution. RSOC is proposed to remove outliers for registering image with simple patterns, low overlapping areas, and large affine transform. In order to eliminate dubious matches, a filtering strategy is designed. The strategy contains two ways spatial order constraints i.e. distance sequence and angle sequence and two decision criteria restrictions, i.e. the stability and accuracy of transformation error. Generally SIFT outperforms previously proposed schemes with respect to robustness but it takes more time whereas SURF take lesser time than SIFT and slightly degrade in robustness. SURF and SIFT is widely used with RSOC in remote sensing, medical imaging, computer vision.