We propose an advanced layering method of integrated dept-position map for real-world object. The depth map of far field object has not contain depth information and impossible to layering the far field objects. Therefore, the position map was rendered from generated high-quality 3D model used deep neural network, it is accurately layering for far field object. However, it has field loss of 3D model depending on the color density, when layering at near field. Therefore, by combining the depth map with the position map, the proposed integrated depth-position map was obtained for accurately layering in far and near field objects.
|