Abstract
Tree-rendering is still an open problem because of high complexity and fractal-like structure consisted in a tree. With two image sequences describing moving paths around a tree and toward a tree, we can synthesize an image sequence that a user walking in the space spanned by the two paths of the two image sequences. Our algorithm works by extracting one image from the source image sequence, and driving pixels in that image with a sequence of displacement field constructed from the original image sequence that corresponds to the desired path specified in advance. Holes are occurred due to this unstable synthesis process, so several methods are presented to synthesize needed information i.e. displacement, color for holes for further synthesis process.