A novel image compression technique employing the self-organised clustering capability of Fuzzy-ART neural network and 2-D runlength encoding is proposed. Initially the image is divided into smaller blocks and the vectors representing the pixels in the blocks are applied to the Fuzzy-ART neural network for classification. The image is then represented by the block codes consisting of the sequence of class indices, and the codebook consisting of the class index and their respective grey levels. Further compression is achieved by 2-D runlength encoding, making use of the repetitions of the class index in the block codes in x and y directions. By controlling the vigilance parameter of Fuzzy-ART, a reasonable compression of the image without sacrificing the image quality can be obtained. From the experimental results, it can be seen that the proposed method of image compression can be used for image communication systems where large compression ratio is required. With the introduction of a new class of Fuzzy-ART network, namely Force Class Fuzzy-ART, hardware implementation of the image compression module is made feasible. This network constrains the maximum number of classes in the output of the network by forcing the new vectors into one of the closest categories.

References

[1]    O. A. Ahmed and M. M. Fahmy, Application of multilayer neural networks to image compression, Proc. IEEE Int. Symposium on Circuits and Systems – ISCAS 97, 2 (1997), pp. 1273–1276.

[2]    F. Ancona and R. Zunino, Concurrent VLSI architecture for vector quantisation Proc. IEEE Int. Symposium on Circuits and Systems – ISCAS 97, 3 (1997), pp. 2076–2079.

[3]    G. A. Carpenter and M. N. Gjaja, Fuzzy-ART choice functions, Proc. World Congress on Neural Networks, 1 (1994), pp. I-713 – I-722.

[4]    G. A. Carpenter, S. Grossberg, and D. B. Rosen, Fuzzy-ART: Fast stable learning and categorisation of analog patterns by an adaptive resonance system, Neural Networks, 4 (1991), pp. 759–771.

[5]    W. Chang, H. S. Soliman, and A. H. Sung, A vector quantisation neural network to compress still monochromatic images, Proc. IEEE Int. Conf. on Neural Networks, 6 (1994), pp. 4163–4168.

[6]    R. D. Dony and S. Haykin, Neural network approaches to image compression, IEEE Proceedings, 83 (1995), pp. 288–303.

[7]    M. Duranton, Image processing by neural networks, IEEE Micro, 16 (1996), pp. 12–19.

[8]    W. C. Fang, B. J. Sheu, and O. T. C. Chen, A neural network based VLSI vector quantizer for real-time image compression, Proc. Data Compression Conference – DCC 91, (1991), pp. 342–351.

[9]    J. Jiang, Neural network technology for image compression, Proc. Int. Broadcasting Convention – IBC 95, (1995), pp. 250–257.

[10]    W. Kou, Digital image compression – Algorithms and standards, Kluwer Academic Publishers, Massachusetts, 1995.

[11]    D. K. Kumar and N. Mahalingam, Nested neural networks for image compression, Proc. IEEE Int. Conf. on Global Connectivity in Energy, Computer, Communication and Control – TENCON 98, 2 (1998), pp. 369–372.

[12]    G. Martinelli, L. P. Ricotti, and G. Marcone, Neural clustering for optimal KLT image compression, IEEE Trans. Signal Processing, 41 (1993), pp. 1737–1739.

[13]    S. A. Rizvi and N. M. Nasrabadi, Lossless image compression using modular differential pulse code modulation, Proc. IEEE Int. Conf. on Image Processing – ICIP 99, 1 (1999), pp. 440–443.

[14]    Valova and Y. Kosugi, Neural network based compression algorithm for gray scale images, Proc. IEEE Int. Joint Symposia on Intelligence and Systems, (1998), pp. 422–428.

[15]    M. A. Wahdan, A. Z. Badr, and T. Mostafa, An efficient image compression technique using variable quadtree resolution and neural networks, Proc. IEEE Sixteenth National Radio Science Conference – NRSC 99, (1999), pp. C16.1–C16/9.

[16]    J. Wu, A. H. Sung, and H. S. Soliman, An ART network with fuzzy control for image data compression, Proc. Third IEEE Conf. on Fuzzy Systems and World Congress on Computational Intelligence, 3 (1994), pp. 1743–1748.