Neural phase microscopy with metasurface optics for real-time and nanoscale quantitative phase imaging.
basic_science · Level V
Where this comes from
- Record sourced from PubMed, PMID 41530131.
- Also identified by DOI 10.1038/s41467-025-68151-z and PMC identifier 12881553.
- Licence recorded as CC BY-NC-ND.
- Because redistribution is not established, this page shows the abstract only. Follow the links below for the full text.
Abstract
Quantitative phase imaging (QPI) enables non-invasive analysis of transparent specimens across biomedicine, materials science, and neuroscience. However, conventional hardware relies on complex architectures with multi-shot acquisition that preclude compact, real-time operation, while phase-retrieval software often yields degraded image quality, is environmentally sensitive, and runs slowly. Here, we introduce a compact, fast, high-resolution QPI platform that addresses these challenges by integrating nanophotonic metasurfaces with artificial intelligence (AI). Our metasurface optics simplifies the optical architecture by replacing bulky optics and modulators, enabling single-shot acquisition and a drastic reduction in form factor. We develop physics-informed AI models that correct optical aberrations, compensate for imperfections in nanofabrication and alignment, and restore nanoscale quantitative phase information in real-time. The system achieves nanoscale resolution better than 840 nm at 74 Hz within a single, thin optical layer. Our unique combination of nanophotonic hardware and AI algorithms advances QPI technology towards portable, precise, real-time phase imaging.