Antique coin image retrieval based on synthetic feature

Multimedia database retrieval is a key technology used to construct a modern digital museum. The antique coin images are the important component parts of multimedia database in digital museum. In this paper, we design an antique coin image retrieval (ACIR) system using synthetic features. The synthetic features include two parts: one is the histogram of centroid distance (HCD) as a coarse feature descriptor and the other is the modified moment invariants (MMI) as a fine feature descriptor. HCD is a contour-based shape feature descriptor with great efficiency when the shapes of objects in two target images have obvious differences. MMI is a region-based feature descriptor which can well distinguish interior details of images. A Synthetic feature matching strategy is proposed to measure the similarity between the query image and database images. The experimental results demonstrate that the algorithm of synthetic feature matching is effective and efficient in ACIR system. And we also believe that our work can dedicate to the construction of digital museum.