Image indexing and retrieval using signature trees

Abstract

Significant research has focused on determining efficient methodologies for effective and speedy retrieval in large image databases. Towards that goal, the first contribution of this paper is an image abstraction technique, called variable-bin allocation (VBA), based on signature bitstrings and a corresponding similarity metric. The signature provides a compact representation of an image based on its color content and yields better retrieval effectiveness than when using classical global color histograms (GCHs) and comparable to the one obtained when using color-coherence vectors (CCVs). More importantly however, the use of VBA signatures allows savings of 75% and 87.5% in storage overhead when compared to GCHs and CCVs, respectively. The second contribution is the improvement upon an access structure, the S-tree, exploring the concept of logical and physical pages and a specialized nearest-neighbor type of algorithm, in order to improve retrieval speed. We compared the S-tree performance when indexing the VBA signatures against the SR-tree indexing GCHs and CCVs, since SR-trees are arguably the most efficient access method for high-dimensional points. Our experimental results, using a large number of images and varying several parameters, have shown that the combination VBA/S-tree outperforms the GCH/SR-tree combination in terms of effectiveness, access speed and size (up to 45%, 25% and 70% respectively). Due to the very highdimensionality of the CCVs their indexing, even using an efficient access structure, the SR-tree, did not seem to be a feasible alternative.

A compact and efficient image retrieval approach based on border/interior pixel classification

“…Lastly, even though we argue that the use of disk-based access structures could be avoided for relatively large image datasets, we believe that the scalability of any CBIR approach is paramount. Hence we should also investigate the use of access structures such as the Signature Tree (e.g., [9]) to speedup query processing.…”

“…The advantages of this log-based representation for histograms are: (1) the comparison of the histograms according to the dLog distance becomes computationally simpler; (2) the histogram can be stored in half of the space of the original representation; (3) as in [9], we can interpret, represent, index and compare histograms as binary signatures.…”

Color Image Retrieval Using Fuzzy Measure Hamming and S-Tree

“…The process of creating the S-tree is based on inserting and splitting the node in the tree [3,6]. The algorithm which creates the S-tree to store the fuzzy signature is as follows: …”

“…So the image retrieval system on the base of the content is developed to extract visual feature to describe the content of image. A number of image retrieval system was built as: QBIC, ADL, Virage, hAlta Vista, SIMPLYcity,… In recent years, the works of query image regarding CBIR, such as: the image retrieval system on the base of the color histogram [4,5], the similarity measure of image on the base of combining color and texture image [7,8], image retrieval technique Variable-Bin Allocation (VBA) using signature and the S-tree [6],… In the approach of the paper will create the fuzzy signature of an image. The content of the paper will aim to efficient query "similar images" in a large database system of digital image.…”

Compressing Large Signature Trees

“…Other improvements for the organization of signatures in tree structures can be found in [TBM02,NM02], which also examine different types of queries (e.g., super-set queries, similarity queries, etc). Extensions of the use of signature indexes to several applications are included in [NTVM02,MNT03] A description of compression schemes that can be applied to tree structures, in general, is given in the [Teuh01]. Experiments in [Teuh01] have focused on the R-tree structure, whereas it also contains a small description regarding the S-tree, which forms the motivation in our approach (for the new contributions of our approach, see the discussion in Section 1.1).…”