Optimizing Ray Tracing with Visual Coherence

We present a method that directs sampling and reconstruction for image generation. This method reduces the cost of the sampling stage, allows a reconstruction of the scene to be progressively reened during the sampling process, and produces a representation of the scene that permits exibility in the generation of images. The method employs a set of region estimators to model and evaluate the samples that are taken of the image. These models describe coherent regions of the image plane. The region estimators are used to direct the sampling process and to provide a description of the scene for reconstruction. The use of region estimators removes the dependence of the reconstruction process upon the sampling process by providing a compact and accurate representation of the sampled data. This representation permits the resolution and lter kernel to be chosen at reconstruction. An algorithm is presented to perform this reconstruction eeciently.