Small-brain neural networks rapidly solve inverse problems with vortex Fourier encoders

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Muminov, Baurzhan, Vuong, Luat T.
Format: Preprint
Published: 2020
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911026653429760
author Muminov, Baurzhan
Vuong, Luat T.
author_facet Muminov, Baurzhan
Vuong, Luat T.
contents We introduce a vortex phase transform with a lenslet-array to accompany shallow, dense, ``small-brain'' neural networks for high-speed and low-light imaging. Our single-shot ptychographic approach exploits the coherent diffraction, compact representation, and edge enhancement of Fourier-tranformed spiral-phase gradients. With vortex spatial encoding, a small brain is trained to deconvolve images at rates 5-20 times faster than those achieved with random encoding schemes, where greater advantages are gained in the presence of noise. Once trained, the small brain reconstructs an object from intensity-only data, solving an inverse mapping without performing iterations on each image and without deep-learning schemes. With this hybrid, optical-digital, vortex Fourier encoded, small-brain scheme, we reconstruct MNIST Fashion objects illuminated with low-light flux (5 nJ/cm$^2$) at a rate of several thousand frames per second on a 15 W central processing unit, two orders of magnitude faster than convolutional neural networks.
format Preprint
id arxiv_https___arxiv_org_abs_2005_07682
institution arXiv
publishDate 2020
record_format arxiv
spellingShingle Small-brain neural networks rapidly solve inverse problems with vortex Fourier encoders
Muminov, Baurzhan
Vuong, Luat T.
Image and Video Processing
Computer Vision and Pattern Recognition
Optics
We introduce a vortex phase transform with a lenslet-array to accompany shallow, dense, ``small-brain'' neural networks for high-speed and low-light imaging. Our single-shot ptychographic approach exploits the coherent diffraction, compact representation, and edge enhancement of Fourier-tranformed spiral-phase gradients. With vortex spatial encoding, a small brain is trained to deconvolve images at rates 5-20 times faster than those achieved with random encoding schemes, where greater advantages are gained in the presence of noise. Once trained, the small brain reconstructs an object from intensity-only data, solving an inverse mapping without performing iterations on each image and without deep-learning schemes. With this hybrid, optical-digital, vortex Fourier encoded, small-brain scheme, we reconstruct MNIST Fashion objects illuminated with low-light flux (5 nJ/cm$^2$) at a rate of several thousand frames per second on a 15 W central processing unit, two orders of magnitude faster than convolutional neural networks.
title Small-brain neural networks rapidly solve inverse problems with vortex Fourier encoders
topic Image and Video Processing
Computer Vision and Pattern Recognition
Optics
url https://arxiv.org/abs/2005.07682