Image Matching Based on Unification

In the image retrieval which is based on content, the two most important operations are extracting the stable image features and matching them. This paper adopts Scale Invariant Feature Transform algorithm (namely SIFT algorithm) to extract image feature points, and then using the theory of unification to perform reliable matching between different views. The features which are extracted from scale-invariant space are invariant to image scale and rotation, and are shown to provide robust matching across a substantial rang of affine distortion, addition of noise, and change in illumination. This paper introduces the unification into the features matching system, the result verify this algorithm has a well adaptability to various conditions, and it improves the match accuracy, at the same time reduces the amount of calculation.