Embedded image compression using a classified multistage VQ in wavelet domain

A wavelet based embedded image coding system, using a classified multistage vector quantization (VQ), is proposed. The classified multistage VQ provides the partial ordering and successive approximation of wavelet coefficient vectors. Using a multiple context model along with a priori information of the vector quantization system, the method produces an optimized bitstream in a rate distortion sense. In addition, coding coefficient vectors rather than single coefficients not only takes advantage of the correlation among coefficients, but it also reduces the significance map overhead leading to coding efficiency. Computer simulations indicate that the proposed scheme is effective and its performance is competitive with state of the art image compression techniques.