Skip to Main Content
In order to develop a system to synthesize arbitrary viewpoint images efficiently, two representative schemes that use the 3-D geometric model to synthesize images have been the focus of attention: model-based rendering and image-based rendering. Image-based rendering is reported to give a better coding performance than the model-based rendering. Although the image-based rendering synthesizes the relatively accurate diffuse reflection component from the camera images at the different viewpoints, it can not synthesize the specular reflection component with sufficient accuracy because the movement of the specular reflection component on the object surface by the viewpoint change is not taken into consideration. Therefore, we propose a method of synthesizing arbitrary viewpoint images by estimating the motion vectors of the specular reflection components for the 3-D model-based synthesized image used as the reference image. Experiments with a carved wooden object show the effectiveness of the proposed method through the coding performance and the subjective evaluation of the synthesized image.