Multiple representations, similarity matching, and results fusion for content-based image retrieval

Abstract

In this paper, we show how the use of multiple content representations and their fusion can improve the performance of content-based image retrieval systems. We consider the case of texture and propose a new algorithm for texture retrieval based on multiple representations and their results fusion. Texture content is modeled using two different models: the well-known autoregressive model and a perceptual model based on perceptual features such as coarseness and directionality. In the case of the perceptual model, two viewpoints are considered: perceptual features are computed based on the original images viewpoint and on the autocovariance function viewpoint (corresponding to original images). So we consider a total of three content representations. The similarity measure used is based on Gower's index of similarity. Simple results of the fusion models are used to merge search results returned by different representations. Experimentations and benchmarking carried out on the well-known Brodatz database show a drastic improvement in search effectiveness with the fused model without necessarily altering their efficiency in an important way.

Effects of motor and sensory nerve transplants on amount and specificity of sciatic nerve regeneration

“…The concept of combining several representations of visual features is not new; many authors have proposed multiple representations of image content (Abbadeni, 2005; Chang & Hsu, 1992; Flickner et al, 1995; French, Watson, Jin, & Martin, 2003; Kailing, Kriegel, & Schonauer, 2004; Urban et al, 2006) and the combination of the results produced by their evaluations (Fagin, 1999; Güntzer, Balke, & Kießling, 2000). In Urban et al (2006) both textual features and color features are used and combined; Kailing et al (2004) use structural and content indexes.…”

“…In Urban et al (2006) both textual features and color features are used and combined; Kailing et al (2004) use structural and content indexes. In Abbadeni (2005), an algorithm for texture retrieval based on multiple representations of the texture and their results fusion is proposed. The QBIC system integrates several visual features such as color and texture described by different methods and several shape features (Flickner et al, 1995).…”

Information retrieval from visual databases using multiple representations and multiple queries

“…In such weighted models, each image j is weighted with its rank in the list of results returned for query i using model M k . Note that the similarity measure used is based on the Gower coefficient of similarity we have developed in our earlier work [6], [7], [8]. Fusion models FusCL and FusComb, both non-weighted and weighted, exploit the chorus effect as well as the dark horse effect while the FusMAX model exploits the skimming effect.…”

A flexible content‐based image retrieval model and a customizable system for the retrieval of shapes

“…In Urban et al (2006) both textual features and color features are used and combined; Kailing et al (2004) use structural and content indexes. In Abbadeni (2005), an algorithm for texture retrieval based on multiple representations of the texture and their results fusion is proposed. The QBIC system integrates several visual features such as color and texture described by different methods and several shape features (Flickner et al, 1995).…”

“…The concept of combining several representations of visual features is not new; many authors have proposed multiple representations of image content (Abbadeni, 2005;Chang & Hsu, 1992;Flickner et al, 1995;French, Watson, Jin, & Martin, 2003;Kailing, Kriegel, & Schonauer, 2004;Urban et al, 2006) and the combination of the results produced by their evaluations (Fagin, 1999;Güntzer, Balke, & Kießling, 2000). In Urban et al (2006) both textual features and color features are used and combined; Kailing et al (2004) use structural and content indexes.…”