Redesigning the Image Itself

Digital images were designed to match machines.
It's time that they were redesigned for people.

	Gigapixel imagery Digital images have been around for so long now that we take them for granted. This project is a fundamental reexamination of the digital image, with the long-term goal of realizing real-time capture and synthesis of gigapixel imagery.
	Temporally adaptive frameless rendering In video compression, temporal coherence can be exploited to achieve higher compression rates than could be achieved frame by frame. We are attempting to exploit that same coherence in interactive rendering.
	Supra-threshold peripheral LOD Most LOD systems make visual manipulations that are quite visible, while basing them on perceptual studies of the point at which objects are only just becoming visual. How should LOD really work above this point?
	Interruptible rendering Here we study a more modest framed approach to exploit temporal coherence. Using progressively refining rendering, we interrupt refinement whenever change to reduce frame times when scenes are dynamic.
	Simplification of massive models Oddly enough, most simplification algorithms cannot simplify extremely large models, because they don't fit in core memory. We are taking a systems based approach to extend rsimp for use with these models.
	Semiautomatic model simplification Automatic simplification methods will not preserve model features of special semantic or application-based importance. We have built a tool that gives the user excellent control over the process of simplification.
	Simplification through refinement This simplification algorithm takes a unique approach, in that it begins with a very crude approximation and refines it. This approach enables extremely fast simplification and scales well to large models.
	Measuring and predicting visual fidelity How can we measure or predict how accurate our imagery is? Which methods are more effective? How are we doing with our simplification algorithms?
	The visual/temporal tradeoff What is the appropriate balance between visual detail (polygons) and delay? We introduce I/O differencing, a new approach to striking this balance.
	Effects of frame rate and delay How do the effects of different mean frame rates on user performance vary by task? Is it important to hold frame rates constant?
	Reducing detail in display peripheries We don't see well at the edges of our field of view. How effective is the reduction of visual detail in the display periphery? How accurately must eye position be tracked?