Image Compression Based on Hierarchical Clustering Vector Quantization - Multimedia and Signal Processing - page 128

From Tables 2 and 3, we can see that compared with traditional algorithms, our
hierarchical clustering VQ framework can achieve higher PSNR performance with
much smaller codebook size. This is attributed to the codebook generation with
smaller vector dimension for structure regions. For this reason, the reconstruction
quality measured by PSNR and SSIM can be improved significantly with much
smaller codebook size at the cost of slight increase in compression rate.

5 Conclusion

In this paper, a new hierarchical clustering VQ framework is proposed. It can generate
the codebooks with different sizes and dimensions according to the different informa-
tion regions, making it more adaptable to the specific regions of an image. By apply-
ing our framework to some basic testing images, the experimental results demonstrate
the superiority of the proposed framework to other traditional algorithms. In the future
work, the traditional FCM algorithm should be optimized to avoid falling into the
local optimum. Besides, we can add more layers and higher dimensional codewords
to reduce the compression bit rate in our framework.

References

1. Gray, R.M.: Vector quantization. IEEE ASSP Magazine, pp. 4-29, (1984)
2. Linde, Y., Buzo, A., Gray, R.M.: An algorithm for vector quantizer design. IEEE Transac-
tions on Communications COM-28, 84-95 (1980)
3. Equitz, W.H.: A new vector quantization clustering algorithm. IEEE Transactions on
Acoustics, Speech and Signal Processing 37(10), 1568-1575 (1989)
4. Somasundaram, K., Vimala, S.: Fast codebook generation for quantization using ordered
pairwise nearest neighbor with multiple merging. In: IEEE International Conference on
Emerging Trends in Electrical and Computer Technology, pp. 581-588 (2011)
5. Akrout, N.M., Prost, R., Goutte, R.: Image compression by vector quantization: a review
focused on codebook generation. Image and Vision Computing 12(10), 627-637 (1994)
6. Kekre, H.B., Sarode, K.T.: Centroid based fast search algorithm for vector quantization.
International Journal of Imaging 1(A08), 73-83 (2008)
7. Kekre, H.B., Sarode, K.T.: Fast codebook search algorithm for vector quantization using
sorting technique. In: ACM International Conference on Advances in Computing, Com-
munication and Control, pp. 23-24, (2009)
8. Vimala, S.: Techniques for generating initial codebook for vector quantization. In: Interna-
tional Conference on Electronics Computer Technology, vol. (4), pp. 201-208 (2011)
9. Samet, H.: The quadtree and related hierarchical data structure. Computer Surveys 16,
187-260 (1984)
10. Yu, P., Venetsanopoulos, A.: Hierarchical multirate vector quantization for image coding.
Signal Processing: Image Communication 4(6), 497-505 (1992)
11. Yu, P., Venetsanopoulos, A.: Hierarchical finite state vector quantization for image cod-
ing. Signal Processing VI: Theories and Applications, 1223-1226 (1992)
12. Klir, G.J., Yuan, B.: Fuzzy sets and fuzzy logic theory and applications, pp. 358-362.
Prentice-Hall Inc., Upper Saddle River (1995)
13. Pal, N.R., Bezdek, J.C.: On cluster validity for the fuzzy c-means model. IEEE Transac-
tions on Fuzzy System 3, 370-372 (1995)
14. Baraldi, A., Blonda, P.: A survey of fuzzy clustering algorithms for pattern recognition-
part I and II. IEEE Transactions on Systems, Man, and Cybernetics-Part B: Cybernet-
ics 29(6) (1999)