Digital image/video compression has been one of the hottest research areas in information technology in the past decade. In this paper

recent advances in image and video compression techniques are introduced. For still image compression

techniques based on human visual perception are emphasized

such as quantiser design in transform domain coding techniques. Techniques using neural network and fractal are introduced too. For video compression

the MPEG-2 technique is emphasized because of its benchmark position for video coding. The evolving MPEG-4 video coding（its general goal and requirement） is mentioned too. Model based coding is introduced as a potential method for MPEG-4. Finally

some current and future applications of digital image/video compresion techniques such as HIDTV are presented.