New technologies for precise light field microscopy in high-resolution neuroimaging: Automated distortion calibration and correction algorithm

Yuan Li#, Yuedi Wang#, Zhi Lu

Tsinghua University, Beijing 100084, China

*Correspondence to: Zhi Lu, PhD, luzhi@tsinghua.edu.cn.

#Both authors contributed equally to this work and share first authorship.

Funding: This work was supported by the Young Scientists Fund of Beijing Natural Science Foundation, No. 4254114 (to ZL), the Natural Science Foundation of China, No. 62575154, the National Postdoctoral Program for Innovative Talent, No. BX20230174, the Postdoctoral Science Foundation of China, No. 2023M741963, the Shuimu Tsinghua Scholar Program, No. 2023SM065, and the Postdoctoral Fellowship Program of CPSF, No. GZC20231305 (to YW).

Objective: Light field microscopy (LFM) enables in vivo neuroimaging with high resolution and thus greatly pushes forward biomedical optical imaging research.

Although continuous development of optical designs andcomputational

reconstructions, inherent image distortion in LFM systems compromises spatial

accuracy and image fidelity, posing a significant barrier to quantitative neuroscience applications. Existing distortion correction methods often rely on intricate,

empirical-dependent calibration procedures or specific calibration targets, limiting their usability and scalability in practical settings.

Methods: We introduce an automated distortion correction algorithm, namely AutoCalibrator-LFM, which utilizes a regular micro-lens-array pattern extracted

from a simple fluorescence acrylic panel to realize distortion characterizations.

Results: Leveraging the intrinsic structure encoded in raw light field data, this method enables precise geometric corrections without the need for external calibration phantoms or manual intervention. The algorithm is validated through extensive simulations and experimental studies involving both static and dynamic biological specimens. This method substantially reduces the alignment artifacts and improves the structural accuracy of reconstructed volumes. Notably, the distortion correction performance is comparable to that of expert manual calibration, while offering a faster and fully automated solution.

Conclusion: AutoCalibrator-LFM makes LFM more accessible and precise and serves as a stepstone for future advanced image-processing algorithms and accurate LFM-based adaptive imaging pipelines in neuroscience.

Key Words: automated calibration; distortion correction; light field microscopy; microlens array; neuroimaging