Lei Tian & Research Team Published in Optica Publishing Group

Recently, Assistant Professor Lei Tian, alongside student researchers Qianwan Yang, Ruipeng Guo, Guorong Hu, Yujia Xue, and Yunzhe Li, published a paper in the Optica Publishing Group. The article, entitled “Wide-field, high-resolution reconstruction in computational multi-aperture miniscope using a Fourier neural network,”  describes their Computational Multi-Aperture Miniscope, SV-FourierNet, which containsa novel solution to current imaging limitations through a computational multi-aperture miniscope, leveraging a microlens array for single-shot, wide FOV, micrometer resolution imaging in a compact and lightweight device.”

This advancement serves as a potentially high-impact change to the fields of biomedical research and beyond, as Tian’s team have proved by enhancing their video constructions of C. Elegans colonies, as well as improving on the imaging of mice brains. The continued improvement of computational imaging has been a long-standing goal of Professor Tian and team, as evidenced in last December’s article on Tian’s computational imaging and microscopy research.

One of the largest issues that has persisted throughout the many evolutions of high-speed, large-scale fluorescence imaging is quality. As lead student author Qianwan Yang explains, “Despite the tremendous progress made in the past decade (e.g. array microscope, light field, and integral imaging techniques), a longstanding tradeoff remains between field-of-view (FOV), resolution, and system complexity. These limitations have impeded the imaging performance in portable applications, such as on-chip microscopy, endoscopy, and in-vivo neural imaging.”

With the SV-FourierNet, the team have forged a new path which utilizes single sensor readings with an unobstructed configuration. In so doing, they are prioritizing the field-of-view (FOV) in a compact form.

“[SV-FourierNet] effectively learn[s] global shift-variant inverse filters in the frequency domain, which not only significantly enhances image quality with higher spatial resolution but also substantially reduces the computational burden compared to spatial domain learning,” Yang says. “To verify the effectiveness of the network, we pioneer the utility of the network’s effective receptive field, demonstrating our network goes beyond basic feature recognition, learning a meaningful reconstruction function aligned with the physical PSFs. Trained purely on synthetic data, SV-FourierNet showcases robust generalization to diverse types of samples, including resolution targets, fluorescent beads, live C. elegans, and weakly scattered brain tissue sections.”

“We believe our Computational Multi-Aperture Miniscope represents a significant advancement in computational microscopy, offering a novel solution for wide-FOV high-resolution imaging in a compact device. We envision it will open exciting new opportunities in a wide range of applications in biomedical research and other areas.” 

Read the Abstract here, and the full article below:

Traditional fluorescence microscopy is constrained by inherent trade-offs among resolution, field of view, and system complexity. To navigate these challenges, we introduce a simple and low-cost computational multi-aperture miniature microscope, utilizing a microlens array for single-shot wide-field, high-resolution imaging. Addressing the challenges posed by extensive view multiplexing and non-local, shift-variant aberrations in this device, we present SV-FourierNet, a multi-channel Fourier neural network. SV-FourierNet facilitates high-resolution image reconstruction across the entire imaging field through its learned global receptive field. We establish a close relationship between the physical spatially varying point-spread functions and the network’s learned effective receptive field. This ensures that SV-FourierNet has effectively encapsulated the spatially varying aberrations in our system and learned a physically meaningful function for image reconstruction. Training of SV-FourierNet is conducted entirely on a physics-based simulator. We showcase wide-field, high-resolution video reconstructions on colonies of freely moving C. elegans and imaging of a mouse brain section. Our computational multi-aperture miniature microscope, augmented with SV-FourierNet, represents a major advancement in computational microscopy and may find broad applications in biomedical research and other fields requiring compact microscopy solutions.

Full Article: https://opg.optica.org/optica/fulltext.cfm?uri=optica-11-6-860&id=552177

View all posts