Real-time retrieval system for heritage images

Abstract

This paper presents a real-time retrieval system of similar images in a large database. The similarity in images is determined by feature matching technique. The Speeded-Up Robust Features (SURF) are computed for all the images in database (pre-computed) and the query image. Along with SURF, the color information of the images is also used for obtaining an efficient similarity among the images. Principal component analysis (PCA) has been carried out to enhance the efficiency of the system, in terms of time and space, which is followed by SR-tree-based multidimensional indexing of the pre-computed image features. The proposed system is experimented on the distributed and centralized computing environments. Experimental results show the performance of the proposed system in the distributed environment in real-time image retrieval process to be a satisfactory format. © 2014 Springer India.