Multiframe superresolution of binary images

Premchandra M. Shankar, Mark A Neifeld

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

We describe a new algorithm for superresolving a binary object from multiple undersampled low-resolution (LR) images that are degraded by diffraction-limited optical blur, detector blur, and additive white Gaussian noise. Two-dimensional distributed data detection (2D4) is an iterative algorithm that employs a message-passing technique for estimating the object pixel likelihoods. We present a novel non-training-based complexity-reduction technique that makes the algorithm suitable even for channels with support size as large as 5 × 5 object pixels. We compare the performance and computational complexity of 2D4 with that of iterative backprojection (IBP). In an imaging system with an optical blur spot matched to the object pixel size, 2 × 2 undersampled measurement pixels, and four LR images, the reconstruction error measured in terms of the number of pixel mismatches for 2D4 is 300 times smaller than that for IBP at a signal-to-noise ratio of 38 dB.

Original languageEnglish (US)
Pages (from-to)1211-1222
Number of pages12
JournalApplied Optics
Volume46
Issue number8
DOIs
StatePublished - Mar 10 2007

Fingerprint

Binary images
Pixels
pixels
image resolution
Image resolution
Message passing
messages
random noise
Imaging systems
Computational complexity
Signal to noise ratio
signal to noise ratios
estimating
Diffraction
Detectors
detectors
diffraction

ASJC Scopus subject areas

  • Atomic and Molecular Physics, and Optics

Cite this

Multiframe superresolution of binary images. / Shankar, Premchandra M.; Neifeld, Mark A.

In: Applied Optics, Vol. 46, No. 8, 10.03.2007, p. 1211-1222.

Research output: Contribution to journalArticle

Shankar, Premchandra M. ; Neifeld, Mark A. / Multiframe superresolution of binary images. In: Applied Optics. 2007 ; Vol. 46, No. 8. pp. 1211-1222.
@article{17c975bd77bb401e9f0ae9180c3560a8,
title = "Multiframe superresolution of binary images",
abstract = "We describe a new algorithm for superresolving a binary object from multiple undersampled low-resolution (LR) images that are degraded by diffraction-limited optical blur, detector blur, and additive white Gaussian noise. Two-dimensional distributed data detection (2D4) is an iterative algorithm that employs a message-passing technique for estimating the object pixel likelihoods. We present a novel non-training-based complexity-reduction technique that makes the algorithm suitable even for channels with support size as large as 5 × 5 object pixels. We compare the performance and computational complexity of 2D4 with that of iterative backprojection (IBP). In an imaging system with an optical blur spot matched to the object pixel size, 2 × 2 undersampled measurement pixels, and four LR images, the reconstruction error measured in terms of the number of pixel mismatches for 2D4 is 300 times smaller than that for IBP at a signal-to-noise ratio of 38 dB.",
author = "Shankar, {Premchandra M.} and Neifeld, {Mark A}",
year = "2007",
month = "3",
day = "10",
doi = "10.1364/AO.46.001211",
language = "English (US)",
volume = "46",
pages = "1211--1222",
journal = "Applied Optics",
issn = "1559-128X",
publisher = "The Optical Society",
number = "8",

}

TY - JOUR

T1 - Multiframe superresolution of binary images

AU - Shankar, Premchandra M.

AU - Neifeld, Mark A

PY - 2007/3/10

Y1 - 2007/3/10

N2 - We describe a new algorithm for superresolving a binary object from multiple undersampled low-resolution (LR) images that are degraded by diffraction-limited optical blur, detector blur, and additive white Gaussian noise. Two-dimensional distributed data detection (2D4) is an iterative algorithm that employs a message-passing technique for estimating the object pixel likelihoods. We present a novel non-training-based complexity-reduction technique that makes the algorithm suitable even for channels with support size as large as 5 × 5 object pixels. We compare the performance and computational complexity of 2D4 with that of iterative backprojection (IBP). In an imaging system with an optical blur spot matched to the object pixel size, 2 × 2 undersampled measurement pixels, and four LR images, the reconstruction error measured in terms of the number of pixel mismatches for 2D4 is 300 times smaller than that for IBP at a signal-to-noise ratio of 38 dB.

AB - We describe a new algorithm for superresolving a binary object from multiple undersampled low-resolution (LR) images that are degraded by diffraction-limited optical blur, detector blur, and additive white Gaussian noise. Two-dimensional distributed data detection (2D4) is an iterative algorithm that employs a message-passing technique for estimating the object pixel likelihoods. We present a novel non-training-based complexity-reduction technique that makes the algorithm suitable even for channels with support size as large as 5 × 5 object pixels. We compare the performance and computational complexity of 2D4 with that of iterative backprojection (IBP). In an imaging system with an optical blur spot matched to the object pixel size, 2 × 2 undersampled measurement pixels, and four LR images, the reconstruction error measured in terms of the number of pixel mismatches for 2D4 is 300 times smaller than that for IBP at a signal-to-noise ratio of 38 dB.

UR - http://www.scopus.com/inward/record.url?scp=34047150007&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34047150007&partnerID=8YFLogxK

U2 - 10.1364/AO.46.001211

DO - 10.1364/AO.46.001211

M3 - Article

C2 - 17318241

AN - SCOPUS:34047150007

VL - 46

SP - 1211

EP - 1222

JO - Applied Optics

JF - Applied Optics

SN - 1559-128X

IS - 8

ER -