Accelerated Backward Warping

In this paper a plane-based backward warping algorithm is proposed to generatenovel views from multiple reference images. First, depth information is employed to reconstructspace planes from individual reference images and calculate the potential occluding relationshipbetween these planes. Then the planes which represent each identical space plane from dierentreference images are compared with each other to decide the one with the best sample rate to bepreserved and used in the later warping period while the other samples are abandoned. While theimage of a novel view is produced, traditional methods in computer graphics, such as visibility testand clipping, are used to process the planes reconstructed. Then the planes processed are projectedonto the desired image from the knowledge on which plane the desired image pixels are warpedfrom can be acquired. Finally, pixels' depth of the desired image is calculated and then a backwardwarping is performed from these pixels to the reference images to obtain their colors. The storagerequirement in the algorithm is small and increases slowly with the number of reference imagesincreases. By combining the strategy of only preserving the best sample parts and the backwardwarping algorithm, the sample problem could be well tackled.