A panoramic video rendering system using a probability mapping method

We have proposed a panoramic video synthesis method that uses a layered probability mapping method to generate high-quality three-dimensional (3D) images. The method's algorithm required a certain amount of synthesis calculation for estimating the presence of objects within an overall scene. Our scheme handles image regions in an adaptive manner by dividing them into an overlap area and a non -overlap area. It uses a layered probability mapping method to process 3D images in the overlap area and projects the images into the non-overlap area in a position estimated from overlap area information. To speed up the synthesis time, we developed a prototype rendering system that makes use of processing performed by personal computers (PCs). Our method calculates the method's algorithm in a parallel manner, thus providing continuous synthetic images. This requires a smaller amount of synthesis calculation than that of the conventional method.