Lensless light-field imaging through complementary segmentation of PSF

Abstract

Light field imaging is applied in fields like virtual reality, 3D reconstruction and medical imaging by capturing spatial and angular information of light to enable more immersive and flexible visual experiences. However, light-field imaging systems often face challenges such as high complexity and limited resolution. In recent years, lensless imaging systems have garnered significant attention for their potential to overcome the limitations of traditional lens-based architectures. Despite their advantages, lensless light-field imaging suffers from notable errors in light-field reconstruction. This paper presents an enhanced Point Spread Function (PSF) segmentation method for lensless light-field imaging, designed to improve the quality of light-field reconstruction. The proposed approach initially divides the PSF into pairs of complementary sub-aperture PSFs, with one PSF targeting the desired sub-aperture area and the other serving as an auxiliary for reconstruction. This allows for the separate reconstruction of sub-aperture images for each area. To assess the light-field reconstruction performance, the imaging quality and angular consistency of various PSF segmentation methods are compared using digital refocusing, epipolar plane images (EPI), depth maps, peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM). Furthermore, the effectiveness of the proposed methodology is validated through experimental results and theoretical analysis. It is demonstrated that lensless light-field imaging using complementary segmentation of the PSF achieves high-quality light-field reconstruction.