Image data processing in the compressed wavelet domain

A new embedded image coding method, the wavelet-difference-reduction algorithm, was introduced by Jun Tian et al. (see Proceedings of the IEEE Data Compression Conference, Snowbird, Utah, 1996). It combines the discrete wavelet transform, differential coding, binary reduction, order bit plane transmission, and adaptive arithmetic coding. It is a simple and efficient way to successively approximate the image source and is very suitable for progressive image transmission. In addition, this algorithm provides a clear geometric structure in the compressed wavelet domain, which enables us to process the image data in compressed form. We discuss three aspects of this problem, namely denoising, speckle reduction, and zooming. These outline some ideas of how to facilitate image data processing in the compressed wavelet domain