Multi-layered geometry image representation of point cloud surfaces

This paper presents a novel approach for representing 3D point cloud surfaces using geometry images. 3D points are converted to images based on spherical polar coordinates, and point cloud surface information such as location, normal and color is stored in two-dimensional images. To create geometry image of point cloud surfaces, we first determine the center of the polar coordinates and convert the Cartesian coordinates [x, y, z] to polar ones [r, ψ, θ], with the coordinate [ψ, θ] encoded as the location and r as the gray value in the image. To represent/reconstruct point cloud surfaces with high complexity, we propose using multi-layered images. From the point cloud surfaces, multi-layered geometry images are created which can represent the original models very accurately. This approach can be used in the compression, editing, simplification, model retrieval and progressive transmission of point cloud surfaces.