Realtime visualization of large images over a thinwire
EC Chang, Chee Keng Yap, TJ Yen · 1997
We investigate realtime techniques for visualization of very large images over a thinwire. We exploit multi-foveated images which have extremely favorable data density, as compared to uniform resolution images. To compensate for the loss of uniform resolution, we now allow the user new dynamic controls over the visualization process. Current servers of large images on the net (e.g., map servers) suffers from a distinctly non-realtime response, its small viewing windows, and visual discontinues in panning/zooming. We have constructed an image server that can overcome these limitations. Our foveation techniques are based on a novel application of wavelets. 1 Introduction We address the realtime visualization of large scale images over a thinwire model of computation. That is, the model comprises an image server and a client viewer connected by a low-bandwidth line. The server may be assumed to have powerful computational resources if necessary; the client has possibly modest resources. ...