Application of Multiwavelets to Image Compression

For evaluating the effectiveness of multiwavelet transform for coding image at low bitrates, an embedded coding of wavelet coefficients has been realized. Two-dimensional image is first prefiltered by Hwee Huat Tan's approach. Then, a new modified embedded zerotree wavelet code (EZW) algorithm with BSA(4/4)* filters is introduced to quantize multiwavelet coefficients. Adaptive arithmetic coding is used to encode the bitstream. Some experimental results show that multiwavelets are effective and worth further studying.