Theoretical analysis of geometry inaccuracy for light field compression

In current light field compression algorithms, geometry information is critical for efficient compression of light field image data. Inaccuracies in the geometry can greatly affect the compression efficiency. In this paper, we present a theoretical framework with which we can analyze and understand the effects of geometry inaccuracy on the compression efficiency of a light field coder. This framework combines ideas from statistical signal processing and rate-distortion theory. We demonstrate using simulations that the theory qualitatively matches the actual results of light field coding algorithms.