Applying a quantum amplification over a system for image feature matching and image recommendation

Abstract

A novel approach for improving the performance of a image feature matching and recommendation system is proposed that is a major continuation of the previous work in [3]. An advanced clustering method is suggested in order to deal with the binary image feature descriptors: a quantum variant of the k-majority algorithm over ORB descriptors. Jaccard-Needham dissimilarity measure is used as a distance measure step of the algorithm. Finally, the Grover's algorithm is used, providing the opportunity for a specific feature search in the database. The paper also provides the main steps in constructing a similar fast search system. The transformation from a classical to a quantum representation algorithm is described. Such approach can be applied in other applications. Both the computational complexity and the verification correctness are also indicated in the paper.