Single image dehazing using deep neural networks

Cameron Hodges, Mohammed Bennamoun, Hossein Rahmani

Research output: Contribution to journalArticle

Abstract

The rapid growth in computer vision applications that are affected by environmental conditions challenge the limitations of existing techniques. This is driving the development of new deep learning based vision techniques that are robust to environmental noise and interference. We propose a novel deep CNN model, which is trained from unmatched images for the purpose of image dehazing. This solution is enabled by the concept of the Siamese network architecture. Using object performance measures of image PSNR and SSIM we are able to demonstrate a quantitative and qualitative improvement in the network dehazing performance. This superior performance is achieved with significantly smaller training datasets than existing methods.

Original languageEnglish
Pages (from-to)70-77
Number of pages8
JournalPattern Recognition Letters
Volume128
DOIs
Publication statusPublished - 1 Dec 2019

Fingerprint

Network performance
Network architecture
Computer vision
Deep learning
Deep neural networks

Cite this

Hodges, Cameron ; Bennamoun, Mohammed ; Rahmani, Hossein. / Single image dehazing using deep neural networks. In: Pattern Recognition Letters. 2019 ; Vol. 128. pp. 70-77.
@article{437a0ab9ab0f4d45a8596812b5b70e97,
title = "Single image dehazing using deep neural networks",
abstract = "The rapid growth in computer vision applications that are affected by environmental conditions challenge the limitations of existing techniques. This is driving the development of new deep learning based vision techniques that are robust to environmental noise and interference. We propose a novel deep CNN model, which is trained from unmatched images for the purpose of image dehazing. This solution is enabled by the concept of the Siamese network architecture. Using object performance measures of image PSNR and SSIM we are able to demonstrate a quantitative and qualitative improvement in the network dehazing performance. This superior performance is achieved with significantly smaller training datasets than existing methods.",
keywords = "Deep learing, Image dehazing, Siamese networks, Single image dehazing",
author = "Cameron Hodges and Mohammed Bennamoun and Hossein Rahmani",
year = "2019",
month = "12",
day = "1",
doi = "10.1016/j.patrec.2019.08.013",
language = "English",
volume = "128",
pages = "70--77",
journal = "Pattern Recogniton Letters",
issn = "0167-8655",
publisher = "Elsevier",

}

Single image dehazing using deep neural networks. / Hodges, Cameron; Bennamoun, Mohammed; Rahmani, Hossein.

In: Pattern Recognition Letters, Vol. 128, 01.12.2019, p. 70-77.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Single image dehazing using deep neural networks

AU - Hodges, Cameron

AU - Bennamoun, Mohammed

AU - Rahmani, Hossein

PY - 2019/12/1

Y1 - 2019/12/1

N2 - The rapid growth in computer vision applications that are affected by environmental conditions challenge the limitations of existing techniques. This is driving the development of new deep learning based vision techniques that are robust to environmental noise and interference. We propose a novel deep CNN model, which is trained from unmatched images for the purpose of image dehazing. This solution is enabled by the concept of the Siamese network architecture. Using object performance measures of image PSNR and SSIM we are able to demonstrate a quantitative and qualitative improvement in the network dehazing performance. This superior performance is achieved with significantly smaller training datasets than existing methods.

AB - The rapid growth in computer vision applications that are affected by environmental conditions challenge the limitations of existing techniques. This is driving the development of new deep learning based vision techniques that are robust to environmental noise and interference. We propose a novel deep CNN model, which is trained from unmatched images for the purpose of image dehazing. This solution is enabled by the concept of the Siamese network architecture. Using object performance measures of image PSNR and SSIM we are able to demonstrate a quantitative and qualitative improvement in the network dehazing performance. This superior performance is achieved with significantly smaller training datasets than existing methods.

KW - Deep learing

KW - Image dehazing

KW - Siamese networks

KW - Single image dehazing

UR - http://www.scopus.com/inward/record.url?scp=85071381558&partnerID=8YFLogxK

U2 - 10.1016/j.patrec.2019.08.013

DO - 10.1016/j.patrec.2019.08.013

M3 - Article

VL - 128

SP - 70

EP - 77

JO - Pattern Recogniton Letters

JF - Pattern Recogniton Letters

SN - 0167-8655

ER -