Non-uniformly sampled aperture-coded light-field microscopy

Light-field microscopy enables three-dimensional volumetric imaging without mechanical scanning, making it highly promising for biomedical research. Conventional microlens-array-based light-field microscopes capture the light-field in a single snapshot but suffer from a trade-off between angular and spatial resolution, which limits spatial resolution. Aperture-coded light-field microscopy achieves full-resolution spatial sampling yet requires multiple measurements, thereby compromising temporal resolution. To overcome this limitation, we propose a non-uniformly sampled aperture-coded light-field microscopy method that exploits the non-uniform energy distribution in the aperture (Fourier) plane. Specifically, dense sampling is applied in the central region and sparse sampling toward the periphery. A single-pixel imaging algorithm is then used to reconstruct the 4D light-field. Experimental results show that our method reduces the number of required encodings by more than half compared to uniformly sampled methods while maintaining high image fidelity within a limited refocusing range. This facilitates the practical adoption of aperture-coded light-field microscopy.