(A) image registration method using inter-feature distance information

Abstract

Image registration is an essential step in many image processing applications that need visual information from multiple images for comparison, integration or analysis. Image Registration means the alignment of images taken at different times, by different sensors or different viewpoints. Recently researchers have introduced image registration techniques using features-based schemes and area-based schemes. Scale Invariant Feature Transform (SIFT) has been the representative method of feature-based scheme for its rotation and scale invariant properties. We introduce a novel feature detection technique which can get scale information using inter-feature distance information at the detection stage. Combining this detection scheme with SIFT description algorithm, the proposed scheme yields less computational load and more simple registration that that of the conventional SIFT. The scheme is examined through extensive experiments on various images and the results show that the proposed scheme exceeds the scheme using SIFT detector in the aspect of match accuracy at the cost of less time consume.