Spectra-polarization microscopic computational imaging for biological applications

This study introduces a microscopic computational imaging system capable of multidimensional spatial, spectral, and polarization detection. The system utilizes four optical encoders and a neural network to reconstruct 36 high-resolution images (1280×960 pixels) across nine spectral bands (400–800 nm), each captured at four polarization angles (0°, 45°, 90°, and 135°), achieving a 1:9 reconstruction ratio. The system performance was evaluated through biological imaging applications, achieving high-resolution visualization of paramecia nuclei, pine pollen surface topography, and mouse muscle fiber birefringence. The integrated design addresses the limitations of conventional single-modality microscopy, providing a versatile tool for biomedical and material analysis.