Selective rendering based on perceptual importance of scene regions

Computer graphics algorithms have for long dealt with simulation of physics: simulation of the geometry of a real-world space, simulation of the light propagation in a real environment and simulation of motor actions with appropriate tracking. Perception principles have subsequently been incorporated into rendering algorithms in order to save rendering computation and produce photo realistic images from a human rather than a machine point of view. Spatial memory tasks are often incorporated in benchmarking processes when assessing fidelity of a VE simulation, since spatial awareness is crucial for human performance efficiency of any task that entails awareness of space. This paper is exploring the exploitation of visual perception principles towards efficient selective rendering techniques. In order to economize on rendering computation, selective rendering guides a high level of detail to appropriate regions of a synthetic scene. Such decisions are based on predictive attention modeling, gaze or even cognitive information. Experimental studies presented will show that it is possible to produce scenes from a human rather than a physics point of view.