Object browsing using the Internet imaging protocol

Abstract This paper builds on the results from the Viseum project where we built an image server/client system to allow browsing of very large images. In the follow-on European ACOHIR project we built systems capable of acquiring colour-calibrated high-resolution views of objects from many positions. A Java viewer allows the user to closely examine objects in a similar way to Quicktime VR but with much higher resolution. The Internet Imaging Protocol is used to allow the viewer to request 64×64 pel tiles on demand to allow fast browsing of the objects in a Web browser. The original image data occupy typically around 200 MBytes yet we can provide almost instantaneous views with zooming and acceptable performance across the Internet or a modem. The approach taken in the Java viewer is modular and easily customised using JavaScript. Caching at both the server and client provide improved performance. This paper shows how the techniques developed for large images have been applied and modified to handle high-resolution object views.