Keynote Lecture - From X to Z via Y: The Fusion of Image Analysis and Synthetic Graphics

Dr John Patterson, University of Glasgow

Abstract

Since the earliest days of raster displays, graphics has owed a grudgingly acknowledged debt to image processing and its handmaiden, digital signal processing (DSP). Classical DSP is developed in one dimension, as is Fourier and wavelet analysis, (so X) but the application of these techniques in digital displays is in two dimensions (so now also Y). Finally modern methods of image analysis from stereo pairs or monocular multiples (movie sequences) using close photogrammetry can give us depth information on everything we can see (so, finally, Z). In order to do this image analysis to the highest accuracy we need to use methods more appropriate to the Computer Aided Design community, while the analysis itself relies on techniques derived from neurophysiology but which have found their most advanced expression from within the Geography community. Although having the appearance of a raid on interdisciplinarity, this fusion of ideas actually reinforces the scholarship of synthetic graphics,while repaying the gesture by picking its pockets. The chief beneficiaries of all this are the digital media industry, who have an insatiable need for capturing 3D models from the real world to put into games and interactive movies, not to mention opening up digital image sequences to the whole range of 3d synthetic graphics techniques.