Presentation
5 March 2021 Generalizable physics-embedded deep learning for recovering complex 3D biology with intensity diffraction tomography
Author Affiliations +
Abstract
We present a physics-embedded deep learning architecture for improving the intensity diffraction tomography’s (IDT) 3D quantitate phase recovery of complex biological samples. Our approach utilizes both convolutional neural networks and IDT’s model-based volumetric reconstructions to provide accurate, physically-constrained predictions of multiple-scattering object volumes. We train and validate this learned IDT pipeline on simulated natural image datasets and show it generalizes well to recovering unseen complex biological specimens from experimental data.
Conference Presentation
© (2021) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Alex C. Matlock and Lei Tian "Generalizable physics-embedded deep learning for recovering complex 3D biology with intensity diffraction tomography", Proc. SPIE 11653, Quantitative Phase Imaging VII, 116530W (5 March 2021); https://doi.org/10.1117/12.2576788
Advertisement
Advertisement
KEYWORDS
Biology

Diffraction

Tomography

Multiplexing

Optical microscopes

Optical tomography

Phase imaging

Back to Top