ABSTRACT
Introduction In this chapter, we compare three optical architectures for compressive imaging: sequential, parallel, and photon sharing. Each of these architectures is analyzed using two different types of projection: (1) principal component (PC) projections and (2) pseudorandom (PR) projections. Both linear and nonlinear reconstruction methods are presented. The performance of each architecture-projection combination is quantified in terms of reconstructed image quality as a function of measurement noise strength. Using a linear reconstruction operator, we show that in all cases of (1), there is a measurement noise level above which compressive imaging is superior to conventional imaging. Normalized by the average object pixel
Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Algorithm Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Architecture Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Sequential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Parallel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Photon Sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Results Using Linear Reconstructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 PC Projections with Linear Reconstruction . . . . . . . . . . . . . . . . . . . . . . . . . 97 PR Projections with Linear Reconstruction . . . . . . . . . . . . . . . . . . . . . . . . .101
