Image based rendering techniques for virtual reality

摘要

In one general aspect, a computer-implemented method can include identifying a plurality of pixel samples included in a layered depth image (LDI) representation of a scene for rendering in a three-dimensional (3D) image in a virtual reality (VR) space, grouping, by a processor, a subset of the plurality of pixel samples into a block of data, including extracting each pixel sample included in the subset of the plurality of pixel samples from the LDI representation of the scene for inclusion in the block of data based on an error metric associated with the respective pixel sample, creating, by the processor, a texture map for a block of data, the texture map being associated with the block of data, storing the block of data and the texture map, and triggering a rendering of the 3D image in the VR space using the block of data and the texture map.