Poster
6 October 2023 Ultrafast imaging using spatiotemporal encoding and deep learning
Chen Zhou, Lidan Zhang, Tunan Xia, Cheng-Yu Wang, Xingjie Ni, Zhiwen Liu
Author Affiliations +
Conference Poster
Abstract
We present a compact, spatiotemporally encoded, deep learning-enabled, single-shot ultrafast imaging system. We simulate the forward light-transport process of an ultrafast event encoded by a random spatiotemporal encoding mask and construct a U-net-based deep learning model to reconstruct the ultrafast event sequences. Trained on simulated ultrafast events consisting of various geometric shapes and handwritten digits with random locations and speeds, the deep learning model can reconstruct multi-frame ultrafast event sequences from simulated single-shot measurements by a normal camera with high reconstruction accuracy and noise tolerance. We also present preliminary fabrication results of spatiotemporal masks. This work provides a simple, cost-effective, single-shot method for studying nonrepetitive ultrafast transient processes.
© (2023) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Chen Zhou, Lidan Zhang, Tunan Xia, Cheng-Yu Wang, Xingjie Ni, and Zhiwen Liu "Ultrafast imaging using spatiotemporal encoding and deep learning", Proc. SPIE PC12681, Ultrafast Nonlinear Imaging and Spectroscopy XI, PC126810G (6 October 2023); https://doi.org/10.1117/12.2679921
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Ultrafast imaging

Deep learning

Device simulation

Ultrafast measurement systems

Ultrafast phenomena

Cameras

Education and training

Back to Top