Fluorescence microscopy is a cornerstone of modern biological research, widely used to reveal cellular structures, molecular interactions, and dynamic life processes. Computational fluorescence microscopy(CFM) has further revolutionized this field by integrating molecular specificity with optical modulation and algorithmic demodulation, enabling high resolution and multidimensional imaging far beyond the limits of conventional wide field microscopy. However, its full potential is still hindered by a long standing challenge: accurate characterization of the imaging system. Traditional approaches either rely on theoretical modeling, which fails to capture the complexity of real optical paths and modulations, or on microsphere based measurement, which suffers from low signal to noise ratio and limited depth. These limitations reduce imaging fidelity and restrict the adaptability of computational fluorescence microscopy in real world applications.
In a new paper published in Light: Science & Applications, a team of scientists, led by Professor Xiaoli Liu from State Key Laboratory of Radio Frequency Heterogeneous Integration, Shenzhen University, Shenzhen 518060, Guangdong, China, Shenzhen Key Laboratory of Intelligent Optical Measurement and Detection, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen 518060, Guangdong, China， and co-workers have developed a sample prior based point spread function (PSF) decoupling method for computational fluorescence microscopy. This strategy integrates optical modulation with computational demodulation, enabling accurate system characterization without the need for sub diffraction particles or fragile theoretical assumptions. Instead, regular biological samples act as modulators to computationally optimize the system PSF, allowing non parametric and adaptive imaging.
The proposed method is centered around a novel strategy of sample prior based PSF decoupling, in which a regular fluorescent sample, is used as an optical modulator to pre-condition the imaging system. By combining its wide field deconvolution result as a computational prior, the system PSF can be accurately optimized without parametric modeling or low signal-to-noise measurements. This approach simultaneously captures system specificity and sample specificity, ensuring faithful recovery of object structures even in complex optical conditions.
Based on this method, they demonstrated a powerful approach that significantly enhances CFM, achieving volumetric imaging comparable to confocal microscopy and multicolor, depth-extended reconstruction across diverse biological tissues. The scientists further explain the advantages of their method:
"Compared with blind deconvolution, which struggles with ill-posed optimization, the strong support of sample priors guarantees accurate PSF decoupling. Our method not only restores fine structures such as multilayer vessels and pollen grains with high contrast, but also enables depth extended and multichannel imaging comparable to confocal microscopy," the authors explain.
"Using a regular sample modulator for PSF decoupling overcomes the issues of low SNR and application limitation using sub-diffraction limited particles, expanding the selection range and diversity of sample references for system characterization." they added.
"Although the CFM used for experimental demonstration is diffraction-limited, the proposed framework of PSF decoupling provides a general strategy of accurate system characterization for diversified imaging modalities. In future research, we will design more advanced computational imaging strategies to relax the dependency on sample priors, enabling flexible adaptation to super-resolution and dynamic live-cell imaging. Ultimately, it provides a promising mechanism and method of system characterization and demodulation for multi-dimensional manipulation and high-performance breakthroughs in CFM." the scientists explain.
