Multiple region-of-interest image coding with embedded watermark

This paper incorporates multiple Region-Of-Interest (ROI) coding and embedded watermark functionalities into SPIHT coding with integer transforms. Each ROI can have an individual shape model and possesses its own priority (reconstruction speed). The priority decides the order of processing for the coefficients at the encoder and decoder. In addition, we modify the refinement stage of SPIHT to embed watermark information into ROI. By inserting the watermark into the zero bits resulting from the left-shift process, the watermark information can be the meaningful image or the useful text for some special applications.