Minimum distortion color image retrieval based on Lloyd-clustered Gauss mixtures

We consider image retrieval based on minimum distortion selection of features of color images modelled by Gauss mixtures. The proposed algorithm retrieves the image in a database having minimum distortion when the query image is encoded by a separate Gauss mixture codebook representing each image in the database. We use Gauss mixture vector quantization (GMVQ) for clustering Gauss mixtures, instead of the conventional expectation-maximization (EM) algorithm. Experimental comparison shows that the simpler GMVQ and the EM algorithms have close Gauss mixture parameters with similar convergence speeds. We also provide a new color-interleaving method, reducing the dimension of feature vectors and the size of covariance matrices, thereby reducing computation. This method shows a slightly better retrieval performance than the usual color-interleaving method in HSV color space. Our proposed minimum distortion image retrieval performs better than probabilistic image retrieval.

[1]  Robert M. Gray.  Gauss mixture vector quantization , 2001, 2001 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.01CH37221).

[2]  Nuno Vasconcelos,et al.  Minimum probability of error image retrieval , 2012, IEEE Transactions on Signal Processing.

[3]  Robert M. Gray,et al.  Robust image compression using Gauss mixture models , 2001 .

[4]  Robert M. Gray,et al.  Image retrieval using color histograms generated by Gauss mixture vector quantization , 2004, Comput. Vis. Image Underst..

[5]  Robert M. Gray,et al.  Classification of features and images using Gauss mixtures with VQ clustering , 2004, Data Compression Conference, 2004. Proceedings. DCC 2004.

[6]  Nuno Vasconcelos,et al.  Bayesian models for visual information retrieval , 2000 .

[7]  Robert M. Gray,et al.  Texture classification based on multiple Gauss mixture vector quantizers , 2002, Proceedings. IEEE International Conference on Multimedia and Expo.

[8]  Geoffrey J. McLachlan,et al.  Finite Mixture Models , 2019, Annual Review of Statistics and Its Application.

[9]  James Ze Wang,et al.  IRM: integrated region matching for image retrieval , 2000, ACM Multimedia.

[10]  Robert M. Gray,et al.  Histogram-based image retrieval using Gauss mixture vector quantization , 2003, 2003 International Conference on Multimedia and Expo. ICME '03. Proceedings (Cat. No.03TH8698).

[11]  Anil K. Jain,et al.  Content-based hierarchical classification of vacation images , 1999, Proceedings IEEE International Conference on Multimedia Computing and Systems.

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