PAVR: High-Resolution Cellular Imaging via a Physics-Aware Volumetric Reconstruction Framework.

Area of Science:

• Cell Biology
• Microscopy
• Deep Learning
• Biophysics

Background:

• Advanced microscopy and deep learning are revolutionizing cell biology.
• High-resolution 3D cell imaging is limited by reconstruction speed, artifacts, and training data requirements.

Purpose of the Study:

• Introduce PAVR, a physics-aware light-field imaging platform.
• Enable fast, end-to-end volumetric reconstruction for cellular imaging.
• Overcome limitations of current 3D imaging techniques.

Main Methods:

• Developed PAVR, integrating single-shot volumetric acquisition with computational reconstruction.
• Trained the system using in silico responses, eliminating the need for external ground-truth data.
• Applied the platform to fixed and live mammalian cells, including cardiomyocytes.

Main Results:

• Demonstrated multicolor volumetric imaging of subcellular organelles.
• Achieved 3D tracking of autofluorescent particles and visualization of organelle dynamics.
• Quantified coupled morphological and functional dynamics in cardiomyocytes under drug treatment.

Conclusions:

• PAVR provides a scalable hardware-software solution for high-throughput 3D cellular imaging.
• Enables quantitative analysis of dynamic cellular systems in research and clinical settings.
• Facilitates sample-independent reconstruction across diverse biological applications.