Generalizable Deep Learning Approach for 3D Particle Imaging using Holographic Microscopy

TL;DR

This paper presents a deep learning method for 3D particle imaging in holographic microscopy that is highly generalizable, fast, and effective across diverse particle types and imaging conditions.

Contribution

A novel deep learning architecture that leverages human perception of diffracted patterns, enabling rapid and generalizable 3D particle analysis from minimal training data.

Findings

Achieves orders of magnitude faster processing speeds.

Performs well across high concentration and noisy datasets.

Handles diverse particle sizes, shapes, and optical properties.

Abstract

Despite its potential for label-free particle diagnostics, holographic microscopy is limited by specialized processing methods that struggle to generalize across diverse settings. We introduce a deep learning architecture leveraging human perception of longitudinal variation of diffracted patterns of particles, which enables highly generalizable analysis of 3D particle information with orders of magnitude improvement in processing speed. Trained with minimal synthetic and real holograms of simple particles, our method demonstrates exceptional performance on various challenging cases including those with high particle concentrations and noises and a wide range of particle sizes, complex shapes, and optical properties exceeding the diversity of the training datasets.