We have used pixels for years. Cameras and displays use them but they aren't the only representation we can use to when manipulating images. They have some disadvantages. They are not scale-free, adding difficulty when we want to display the same picture on different devices. They are very localised samples, only weakly really related to the underlying optical image. Heavy users in the visual effects industry can have problems with quality after multiple sequential operations.
We have been investigating other ways of representing images, more in keeping with the way synthetic images are created. We will explain our experiments and also show a professional-quality movie made using our latest method.
Download a PDF version of the slides for this presentation. This may take a while as it is a large (13 MB) file.
About the speaker:
Phil Willis is Professor of Computing at the University of Bath, UK. His research interests are in computer graphics.
He has led the Department of Mathematical Sciences and also the Department of Computer Science, which he founded. He has been the chair of the Eurographics Association and became the first formal joint member of EG and ACM SIGGRAPH following the signing of a new cooperation agreement between them.
He directs the nationally-funded Centre for Digital Entertainment. This is a doctoral training centre with 50 graduate students supervised by 35 faculty at two universities. Unusually, every research student spends three years following their research programme in a company. Companies involved include Electronic Arts, Double Negative Visual Effects, Sony, the BBC and more than 30 others.
Department of Computer Science
University of Bath
Bath BA2 7AY