Is there any future in image compression research?

The evolution of visual communication systems show a dangerous lack of use of new research as well as the enhancement and adaptation of old schemes. So, almost classical time prediction and spatial DC T still survive against new proposals. In parallel, new ideas coming from the images analysis and computer vision areas can increase the speedperformance rate of that schemes, although not modifying their underlying principles. On the other hand, model-based approaches seem to offer interesting alternatives, but they stem from computer graphics, using different paradigms. Finally, several private-based alternatives are been proposed to protect proprietary visual information in open domains like the Internet. Looking to all these facts, it seems that future research, if any, should be addressed to the joint real-time graphics-video transmission.