Toroidal Adaptive Intensity and Spectrum Updating Image Reconstruction for Fourier Ptychographic Microscopy

Fourier Ptychographic Microscopy (FPM) is a popular computational imaging technique that reconstructs high-resolution and wide-field images from multiple low-resolution ones acquired with varying illumination angles, overcoming the optical diffraction limit. This study proposes an image reconstruction enhancement to FPM, named toroidal adaptive intensity and spectrum updating. It employs an image intensity update strategy that replaces low-resolution image intensities with varying fractional power based on illumination angles and image frequency content, which in turn adaptively determines how the reconstruction updates in the frequency domain. It consists of three key components: (a) determining two fractional powers for intensity updating, one of which is decided by illumination angles, and the other by wavelet transform frequency content; (b) obtaining a weighted sum of the two fractional powers; (c) an adaptive image intensity and spectrum updating scheme using the fractional power. Experimental validation was conducted on the USAF 1951 chart and 10 biological specimen slices using a Pentax microscope with a 15x15 LED array for illumination. It demonstrates reduced artifacts and improved local contrast compared to conventional methods.