Search papers, labs, and topics across Lattice.
This study introduces a physics-informed generative adversarial network (GAN) that enhances cross-modality super-resolution in fluorescence microscopy by integrating microscope-specific point spread function (PSF) information into the training process. By evaluating both simulated and experimentally measured PSFs with a limited dataset of TOM20-labeled mitochondria, the model significantly outperformed traditional data-driven approaches in terms of structural fidelity and alignment with STED references. The findings underscore the importance of incorporating optical priors to enhance the physical plausibility of generative models in microscopy applications.
Incorporating optical priors into generative models can dramatically enhance the fidelity of super-resolution fluorescence microscopy images.
Cross-modality image translation offers a route to super-resolution fluorescence microscopy from low-resolution images while reducing phototoxicity and instrumentation demands. However, purely data-driven models can produce visually plausible outputs that are inconsistent with optical image formation. Here, we propose a physics-informed generative adversarial network for confocal-to-STED image translation that incorporates microscope-specific point spread function information into the training objective. Simulated and experimentally measured PSFs were evaluated using a limited paired confocal-STED dataset of TOM20-labeled mitochondria in human primary M2 macrophages acquired across different experimental days. Performance was assessed using reference-based and non-reference-based image-quality metrics, together with complementary frequency- and distribution-sensitive analyses. The no-reference metrics probed physics-relevant image properties, including spatial-frequency content, contrast, and signal-to-noise behavior. PSF-guided models improved structural fidelity, reduced local deviations, and achieved closer agreement with STED references than non-PSF baselines, particularly in frequency-domain analyses. These results demonstrate that optical priors can improve the structural fidelity and physical plausibility of generative microscopy models for cross-modality super-resolution imaging.
