An Automatic Focusing Method for Cell Holographic Images in Microfluidic Lensless Imaging Systems

This paper proposed an autofocus method for reconstructing cell hologram images captured in a lensless imaging system with microfluidics. The reconstructed evaluation functions are first analyzed deeply because of single-peak property, robustness, and search sensitivity. Then the golden segmentation was used to search in two steps, with a rough search in a large area and a refined search in a small area. When the search width is 1.6mm, and the search accuracy is 0.036mm, the number of searches is reduced by 77% compared with the traversal method and 57% compared with the hill climbing method. The SSIM and PSNR improved by 9.22% and 9.51% compared with the climbing method after reconstruction.