JPEG 2000 is a new image coding system that delivers superior compression performance and provides many advanced features in scalability, flexibility, and system functionalities that outperform all previous standards. It brings exciting possibilities to many imaging applications such as the Internet, wireless, security, and digital cinema. This paper gives an overview of this triumph in innovations and teamwork. It gives brief introductions to the four new parts that are under development by the JPEG committee. It attempts to summarize the lessons learned from the JPEG 2000 development experience and draw some conclusions on the success factors of this important standard.

The widespread adoption of the Internet as a new paradigm for delivering client/server imaging applications has created exciting opportunities for digital imaging. This paper presents several examples of how Hewlett-Packard is building collaborative relationships with industry and universities to help meet the challenge. To encourage industries to collaborate on development of open industry standards, Hewlett-Packard has worked with industries to develop new image file formats and to contribute to imaging standardization activities. To encourage interdisciplinary collaborations in imaging research and education, Hewlett-Packard is supporting the development of university curricula that combine the disciplines of psychology, neuroscience, computer science and electrical engineering into a coherent engineering curriculum. To stimulate the development of new image processing methods, Hewlett-Packard Laboratories is investigating new approaches to image processing, such as Occam filtering and geometry driven partial differential equations for imaging.

This paper presents a multiprocessor system architecture and the scheduling policies of the processors which can off-load a host computer in the image processing of office documents. The objective of this system is to provide extensive image processing capabilities at minimum cost and with an acceptable interactive response time. Multiprocessor system architecture, image partitioning strategies, image processing algorithm characteristics, performance modeling, and processor scheduling policies are investigated. A prototype multiprocessor image processing system has been built at the IBM San Jose Research Laboratory.