بهينه سازي بانك فيلتر در كدينگ زير باند براساس تبادل گستردگي در مشخصه حوزه فركانس و مكان

For image data transferring and saving compression is inevitable; so the various compression methods have been developed. Among various methods the mapping of data image to uncorrelated coefficients and compression a large fraction of the average energy in a relatively few components is very important. The subband coding (SBC) is a powerful method to achieve this aim which performs better than the traditional orthogonal transform. In the subband coding, the trade off between the resolution in space and frequency domain leads to indecision in choosing of the,filter hank. Although the choice of sharp filter leads to a better frequency decomposition but the impulse response will spread out and this causes a ringing in the edgeʹs area of the image; known as mosquito noise. In this paper it is illustrated that in the hierarchical decomposition; only at the primary levels the sharp filters is needed and for the next levels the taps of the filters could be reduced. This method preserves the mean square error of the reconstructed image and performs better for the sharpness of the edges. Furthermore, by this method the amount of the calculation would be extremely reduced.