Optimization of illumination for Fourier ptychographic microscopic imaging based on regional energy sensitivity and array sub-region delineation

Abstract Fourier ptychographic microscopy (FPM) achieves high-resolution reconstruction by stitching multi-angle illuminated images in the frequency domain, but its traditional full sampling mode requires activating a large number of LED sources, leading to data redundancy and excessive time costs. To address these issues, this paper proposes an illumination optimization strategy based on regional energy sensitivity (RES) and LED array sub-region division, termed RES-FPM, to improve the imaging efficiency of FPM. The method first constructs a frequency-domain sampling window based on the initial high-resolution spectrum map and generates the spatial response map corresponding to each LED through inverse Fourier transform. It then calculates the degree of influence of each LED on the image’s regional energy (RES score). To avoid spatial non-uniformity caused by direct global ranking-based LED selection, this work further divides the LED array into spatial sub-regions. Within each sub-region, several representative LEDs are selected based on their local energy contribution, achieving a spatial-energy-balanced co-optimization. Experimental results demonstrate that, while maintaining spectral coverage and image reconstruction quality, RES-FPM reduces the required image acquisitions by ∼54.4%, yet preserves reconstruction performance comparable to traditional FPM sampling. This significantly lowers computational costs for both data acquisition and reconstruction while improving efficiency. Additionally, it exhibits strong robustness and interpretability when dealing with complex sample structural variations.