06

2026

-

03

Neural Phase Metamicroscopy for Real-Time Nanoscale Quantitative Phase Imaging

Author:


The teams led by Gordon Wetzstein and Gun-Yeal Lee at Stanford University, and Yoonchan Jeong and Changhyun Kim at Seoul National University, have developed a compact, fast, and high-resolution quantitative phase imaging (QPI) platform that combines nano-photonics metasurfaces with artificial intelligence (AI). By replacing bulky optical components and modulators, the authors’ metasurface-based optical system simplifies the optical architecture, enabling single-shot acquisition and significantly reducing the overall size. The authors have also developed an AI model grounded in physical principles, capable of correcting optical aberrations, compensating for imperfections in nanofabrication and alignment, and recovering nanoscale quantitative phase information in real time. This system achieves nanoscale resolution better than 840 nm and a sampling rate of 74 Hz within a single, thin optical layer. The authors’ unique combination of nano-photonics hardware and AI algorithms is driving QPI technology toward portable, high-precision, real-time phase imaging applications.

The research findings were published in Nature Communications on January 13, 2026, under the title “Neural phase microscopy with metasurface optics for real-time and nanoscale quantitative phase imaging.”

Figure 1: Schematic Diagram of the QPI Microscope Optical System Configuration

Figure 2: Design and Characterization of a Metasurface with Complex Amplitude

Figure 3: Schematic Diagram of the Neural Phase Microscopy Procedure

Figure 4: Performance Evaluation of the Phase Retrieval Model via Baseline Comparison

Figure 5: Experimental results of the metasurface neural phase microscope

Source: Optical World