Home
Registration
Super-Resolution
Bacteria Identification
Doppler analysis
MSL segmentation
CGMM
X-ray classification
Cervigrams
Image Clustering
Video Indexing
Object tracking
Gmm&Edge

 

Overview

In recent years there has been a growing interest in developing effective methods for content-based image retrieval (CBIR). Image clustering and categorization is a means for high-level description of image content. The goal is to find a mapping of the archive images into classes (clusters) such that the set of classes provide essentially the same information about the image archive as the entire image-set collection. The generated classes provide a concise summarization and visualization of the image content that can be used for different tasks related to image database management. Image clustering enables the implementation of efficient retrieval algorithms and the creation of a user-friendly interface to the database.

A common approach to image clustering involves addressing the following issues:

1.      Image features how to represent the image.

2.      Organization of feature data how to organize the data.

3.      Classifier how to classify an image to a certain cluster.

In our work we propose a new method for unsupervised image clustering. We use probabilistic continuous Gaussian Mixture models (GMM), for the image representation, and information theoretic principles for image classification and database organization.

Examples for images and their related GMM representation

ellipsoid represents the support, mean color and spatial layout of a particular Gaussian in the image plane.

For an overview of the method see the following slides.
For a detailed explanation read the Thesis.

 

Publications

bullet

J. Goldberger, S. Gordon and H. Greenspan.
"
Unsupervised image-set clustering using an information theoretic framework."
IEEE Trans. on Image Processing, 15(2):449-458, 2006.
[pdf]

bullet

S. Gordon, H. Greenspan and J. Goldberger
"Applying the Information Bottleneck principle to unsupervised clustering of discrete and continuous image representations", In Proc. of  International Conference on Computer Vision, pages 370-377, Nice, France, 2003.
[ps]

bullet

J. Goldberger, S. Gordon and H. Greenspan
"An efficient image similarity measure based on approximations of KL-divergence between two Gaussian Mixtures", In Proc. of  International Conference on Computer Vision, pages 487-493, Nice, France, 2003.
[ps]

bullet

J. Goldberger, H. Greenspan, and S. Gordon.
"Unsupervised image clustering using the Information Bottleneck method", In DAGM 2002.
[pdf]

bullet

H. Greenspan, S.Gordon and J. Goldberger.
"Probabilistic models for generating, modeling and matching image categories", In Proc. of the International Conference on Pattern Recognition, Quebec, August 2002.
[pdf] 

bullet

H. Greenspan, J.  Goldberger and L.Ridel.
"A continuous probabilistic framework for image matching", Journal of Computer Vision and Image Understanding.  84:384-406, 2001.
[pdf]

 

horizontal rule

Home Registration Super-Resolution Bacteria Identification Doppler analysis MSL segmentation CGMM X-ray classification Cervigrams Image Clustering Video Indexing Object tracking Gmm&Edge

For problems or questions regarding this web page contact hayit@eng.tau.ac.il.
Last updated: 01/07/08.