Signal model and granular-noise analysis of computational image reconstruction for curved integral imaging systems

摘要

In this paper, we propose an improved analysis on the signal property of curved computational integral imaging reconstruction (C-CIIR). In the proposed model and analysis, we explain a general analysis of computational integral imaging by introducing a curvature effect that is obtained by the additional use of a large-aperture (LA) lens. Based on the proposed signal model in C-CIIR, we analyze the characteristics of the granular noise (GN) and conduct preliminary experiments to show the feasibility of our model. Experimental results indicate that the GN caused by the nonuniform overlapping gets reduced and that the GN is diminished as the focal length of the additional LA lens used decreases in C-CIIR. Also, the proposed model and analysis are considered to be generalized versions of the signal model and analysis of the previous computational integral imaging systems.