Hybrid approach for still image compression based on fractal approximation and vector quantization

This paper presents a hybrid approach to image compression based on vector quantization (VQ) and fractal approximation. The low frequency components of an input image are approximated by VQ and its residual is coded by fractal mapping. Instead of using indirectly the gray patterns of an original image with contraction mapping for a domain pool like in the conventional fractal coding algorithms, this fractal coding method firstly employs an image approximated by transform VQ (TVQ) and then is decimated as a domain pool. With the proposed algorithm, the constraint of contraction mapping is not required, fractal approximation that uses the self-similarity of gray patterns works on the approximated image. Also, in order to improve the encoding step and to reduce the complexity of the codec, the authors introduce the orthogonalization of the domain pool. For designing the codebook of the TVQ, the Lind Buzo and Gray (LBG) algorithm is used. Computer simulations with several test images show that the proposed method yields better performance than the conventional fractal coding methods for encoding still images