Greet Frederix

Finding salient regions in images: nonparametric clustering for image segmentation and grouping

A major problem in content-based image retrieval (CBIR) is the unsupervised identification of perceptually salient regions in images. We contend that this problem can be tackled by mapping the pixels into various feature-spaces, whereupon they are subjected to a grouping algorithm. In this paper we develop a robust and versatile nonparametric clustering algorithm that is able to handle the unbalanced and highly irregular clusters encountered in such CBIR applications. The strength of our approach lies not so much in the clustering itself, but rather in the definition and use of two cluster-validity indices that are independent of the cluster topology. By combining them, an optimal clustering can be identified, and experiments confirm that the associated clusters do, indeed, correspond to perceptually salient image regions.

Panoramic, adaptive and reconfigurable interface for similarity search

We outline the architecture of a content-based image retrieval (CBIR)-interface that offers the user a graphical tool to create new features by showing (as opposed to telling!) the system what he means. It allows him to interactively classify images by dragging and dropping them into different piles and instructing the interface to come up with features that can mimic this classification. We show how logistic regression and Sammon projection can be used to supervise this search mode.

Image Segmentation by Nonparametric Clustering Based on the Kolmogorov-Smirnov Distance

In this paper we introduce a non-parametric clustering algorithm for 1-dimensional data. The procedure looks for the simplest (i.e. smoothest) density that is still compatible with the data. Compatibility is given a precise meaning in terms of the Kolmogorov-Smirnov statistic. After discussing experimental results for colour segmentation, we outline how this proposed algorithm can be extended to higher dimensions.

Show Me What You Mean! Pariss: A CBIR-Interface that Learns by Example

We outline the architecture of a CBIR-interface that allows the user to interactively classify images by dragging and dropping them into different piles and instructing the interface to come up with features that can mimic this classification. Logistic regression and Sammon projection are used to support this search mode.

Finding regions of interest for content-extraction

A major problem in content based image retrieval (CBIR) is the unsupervised identification of perceptually salient regions in images. We contend that this problem can be tackled by mapping the pixels into various feature-spaces, whereupon they are subjected to a grouping algorithm. In this paper, we develop a robust and versatile non-parametric clustering algorithm that is able to handle the unbalanced and highly irregular clusters encountered in such CBIR applications. The strength of our approach lies not so much in the clustering itself, but rather in the definition and use of two cluster-validity indices that are independent of the cluster topology. By combining them, an optimal clustering can be identified, and experiments confirm that the associated clusters do, indeed, correspond to perceptually salient image regions.

Shape-Invariant cluster validity indices

This paper discusses two cluster validity indices that quantify the quality of a putative clustering in terms of label-homogeneity and connectivity. Because the indices are defined in terms of local data-density, they do not favour spherical or ellipsoidal clusters as other validity indices tend to do. A statistics-based decision framework is outlined that uses these indices to decide on the correct number of clusters.

Image segmentation based on statistically principled clustering

A statistically principled approach to 1-dimensional clustering was introduced by Pauwels abd Frederix (2000). In this approach clustering is achieved by finding the smoothest density that is statistically compatible with the observed data. In the current contribution we propose two solutions for the optimisation problem that is at the heart of this algorithm. The first solution is based on spline-functions, while the second hinges on an expansion of the density in terms of Gaussians. The latter is reminiscent of mixture-models but fundamentally different in its interpretation. Finally, we argue that 1-dimensional histogram segmentation yields a powerful local nonparametric cluster-validity criterion that can be used to check the quality of proposed clusterings in higher dimensions.

Cluster-based segmentation of natural scenes

In cluster-based segmentation pixels are mapped into various feature spaces whereupon they are subjected to a grouping algorithm. In this paper we develop a robust and versatile non-parametric clustering algorithm that is able to handle the unbalanced and irregular clusters encountered in such segmentation applications. The strength of our approach lies in the definition and use of two cluster validity indices that are independent of the cluster topology. By combining them, an excellent clustering can be identified, and experiments confirm that the associated clusters do indeed correspond to perceptually salient image regions.

Automatic Interpretation Based on Robust Segmentation and Shape-Extraction

We report on preliminary but promising experiments that attempt to get automatic annotation of (parts of) real images by using non-parametric clustering to identify salient regions, followed by a limb-characterization algorithm applied to the contours of the regions.

Content-based image retrieval as a tool for image understanding

We propose a robust non-parametric algorithm for the segmentation of 1D histograms. This algorithm is based on the empirical distribution function of the data and the resulting partitioning of the histogram can be used to identify salient regions in images. Furthermore, we report on experiments where these extracted salient regions are used for region-based similarity matching.