An improved EM algorithm for content based image retrieval

Content Based Image Retrieval (CBIR) mainly contains two phases: first, to represent an image; second, to measure the dissimilarity between two images. Expectation-Maximization (EM) is a popular algorithm for clustering Gauss mixtures for the image representation, but the greedy nature of EM make it hard to get an optimal model for CBIR. In this paper, we introduce an improved EM algorithm for clustering Gaussian Mixtures (GMs) to represent an image, instead of regular EM algorithm. We also use different dissimilarity measures for different queries according to their statistics features. Experiments show this approach can greatly improve the performance of CBIR.

[1]  Brendan J. Frey,et al.  A comparison of algorithms for inference and learning in probabilistic graphical models , 2005, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[2]  Anil K. Jain,et al.  Image classification for content-based indexing , 2001, IEEE Trans. Image Process..

[3]  Allen Gersho,et al.  Vector quantization and signal compression , 1991, The Kluwer international series in engineering and computer science.

[4]  Spyros Liapis,et al.  Color and texture image retrieval using chromaticity histograms and wavelet frames , 2004, IEEE Transactions on Multimedia.

[5]  Gérard Govaert,et al.  An EM algorithm for the block mixture model , 2005, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[6]  Jin Zhang,et al.  Weighted expectation maximization reconstruction algorithms for thermoacoustic tomography , 2005, IEEE Transactions on Medical Imaging.

[7]  Leonidas J. Guibas,et al.  A metric for distributions with applications to image databases , 1998, Sixth International Conference on Computer Vision (IEEE Cat. No.98CH36271).

[8]  Nuno Vasconcelos,et al.  On the efficient evaluation of probabilistic similarity functions for image retrieval , 2004, IEEE Transactions on Information Theory.

[9]  Gunther Wyszecki,et al.  Color Science: Concepts and Methods, Quantitative Data and Formulae, 2nd Edition , 2000 .

[10]  Jianping Fan,et al.  Statistical modeling and conceptualization of natural images , 2005, Pattern Recognit..

[11]  幹男 山本.  “IEEE Trans. on Medical Imaging”誌と本誌 , 1983 .

[12]  Robert M. Gray,et al.  Minimum distortion color image retrieval based on Lloyd-clustered Gauss mixtures , 2005, Data Compression Conference.

[13]  J. Cohen,et al.  Color Science: Concepts and Methods, Quantitative Data and Formulas , 1968 .