Deep-learning-augmented computational miniature mesoscope

Fluorescence microscopy is essential to study biological structures and dynamics. However, existing systems suffer from a trade-off between field of view (FOV), resolution, and system complexity, and thus cannot fulfill the emerging need for miniaturized platforms providing micron-scale resolution across centimeter-scale FOVs. To overcome this challenge, we developed a computational miniature mesoscope (CM2) that exploits a computational imaging strategy to enable single-shot, 3D high-resolution imaging across a wide FOV in a miniaturized platform. Here, we present CM2 V2, which significantly advances both the hardware and computation. We complement the 3 x 3 microlens array with a hybrid emission filter that improves the imaging contrast by 5x, and design a 3D-printed free-form collimator for the LED illuminator that improves the excitation efficiency by 3x. To enable high-resolution reconstruction across a large volume, we develop an accurate and efficient 3D linear shift-variant (LSV) model to characterize spatially varying aberrations. We then train a multimodule deep learning model called CM(2)Net, using only the 3D-LSV simulator. We quantify the detection performance and localization accuracy of CM(2)Net to reconstruct fluorescent emitters under different conditions in simulation. We then show that CM(2)Net generalizes well to experiments and achieves accurate 3D reconstruction across a similar to 7-mm FOV and 800-mu m depth, and provides similar to 6-mu m lateral and similar to 25-mu m axial resolution. This provides an similar to 8x better axial resolution and similar to 1400x faster speed compared to the previous model-based algorithm. We anticipate this simple, low-cost computational miniature imaging system will be useful for many large-scale 3D fluorescence imaging applications. (C) 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement

Automated summary

A computational miniature mesoscope (CM2) was developed to enable single-shot, 3D high-resolution imaging across a wide field of view on a miniaturized platform. By improving hardware and computation, including a hybrid emission filter and a 3D-printed collimator for LED illuminator, along with the development of a 3D linear shift-variant model and a deep learning model, accurate and efficient 3D reconstruction was achieved. The CM2Net model demonstrated superior axial resolution and speed compared to previous algorithms, making it a promising tool for large-scale 3D fluorescence imaging applications.