Prior-free depth-of-field extension capability of Fourier ptychographic microscopy via various optimizers

Fourier ptychographic microscopy (FPM) has emerged as a powerful imaging technique, offering a high space bandwidth product that theoretically resolves the trade-off between spatial resolution and field of view. Recently, a novel, to our knowledge, computational framework, termed feature domain Fourier ptychographic microscopy (FD-FPM), has received wide attention because of its ability to extend the depth-of-field of FPM. In order to further clarify the depth-of-field extension capability of FD-FPM, our paper expounds on the principle that FPM based on feature extraction can reconstruct defocused data beyond the depth-of-field of the objective lens. We also analyze the differences in the iterative reconstruction of the defocused data via various optimizers within this framework. By incorporating advanced optimizers into the reconstruction process, a reconstruction with a resolution of 1.5 µm can be achieved with an object-side defocus distance of 200 µm. We have also demonstrated the capability of this computational framework to reconstruct defocused data on biological samples.