Towards Content Independent No-reference Image Quality Assessment Using Deep Learning

Tan Lu; Ann Dooms

doi:10.1109/ICIVC47709.2019.8981378

Towards Content Independent No-reference Image Quality Assessment Using Deep Learning

Tan Lu
, Ann Dooms

Research output: Chapter in Book/Report/Conference proceeding › Conference paper

7 Citations (Scopus)

Abstract

The study of image quality assessment (IQA) is divided on natural scene and document images which are processed using different models and quality metrics. This casts challenges for the development of content-independent no-reference (NR) IQA models which can operate on different types of images without requiring information regarding the content of the images. In this paper we propose a unified no-reference image quality assessment (UIQA) model using a deep learning approach, where a generalization of NR IQA across natural scene and document images is achieved using a deep convolutional neural network (DCNN). Without having to discriminate the type of the images, the proposed model can assess the quality of natural scene and document images in a blind and uniform manner. Testing results on two benchmarking datasets demonstrate that the proposed model achieves promising performances competitive with the state-of-the-art simultaneously on natural scene and document images.

Original language	English
Title of host publication	2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC)
Place of Publication	Xiamen, China
Publisher	IEEE
Pages	276-280
Number of pages	5
Volume	July-2019
Edition	July-2019
ISBN (Electronic)	978-1-7281-2325-7
ISBN (Print)	978-1-7281-2326-4
DOIs	https://doi.org/10.1109/ICIVC47709.2019.8981378
Publication status	Published - Jul 2019
Event	2019 4th IEEE International Conference on Image, Vision and Computing - Huaqiao University, Xiamen, China Duration: 5 Jul 2019 → 7 Jul 2019 http://www.icivc.org/icivc19.html

Publication series

Name	Proceedings of 2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC)
Publisher	IEEE
Number	July-2019

Conference

Conference	2019 4th IEEE International Conference on Image, Vision and Computing
Abbreviated title	ICIVC
Country/Territory	China
City	Xiamen
Period	5/07/19 → 7/07/19
Internet address	http://www.icivc.org/icivc19.html

Keywords

no-reference image quality assessment
perceptual score
OCR accuracy
deep convolutional neural network
transfer learning

Access to Document

10.1109/ICIVC47709.2019.8981378

Handle.net

Persistent link

Cite this

Lu, T., & Dooms, A. (2019). Towards Content Independent No-reference Image Quality Assessment Using Deep Learning. In 2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC) (July-2019 ed., Vol. July-2019, pp. 276-280). Article 8981378 (Proceedings of 2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC); No. July-2019). IEEE. https://doi.org/10.1109/ICIVC47709.2019.8981378

@inproceedings{f2373afe3d1347628ae74766aa7a48aa,

title = "Towards Content Independent No-reference Image Quality Assessment Using Deep Learning",

abstract = "The study of image quality assessment (IQA) is divided on natural scene and document images which are processed using different models and quality metrics. This casts challenges for the development of content-independent no-reference (NR) IQA models which can operate on different types of images without requiring information regarding the content of the images. In this paper we propose a unified no-reference image quality assessment (UIQA) model using a deep learning approach, where a generalization of NR IQA across natural scene and document images is achieved using a deep convolutional neural network (DCNN). Without having to discriminate the type of the images, the proposed model can assess the quality of natural scene and document images in a blind and uniform manner. Testing results on two benchmarking datasets demonstrate that the proposed model achieves promising performances competitive with the state-of-the-art simultaneously on natural scene and document images.",

keywords = "no-reference image quality assessment, perceptual score, OCR accuracy, deep convolutional neural network, transfer learning",

author = "Tan Lu and Ann Dooms",

year = "2019",

month = jul,

doi = "10.1109/ICIVC47709.2019.8981378",

language = "English",

isbn = "978-1-7281-2326-4",

volume = "July-2019",

series = "Proceedings of 2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC)",

publisher = "IEEE",

number = "July-2019",

pages = "276--280",

booktitle = "2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC)",

edition = "July-2019",

note = "2019 4th IEEE International Conference on Image, Vision and Computing ; Conference date: 05-07-2019 Through 07-07-2019",

url = "http://www.icivc.org/icivc19.html",

}

Lu, T & Dooms, A 2019, Towards Content Independent No-reference Image Quality Assessment Using Deep Learning. in 2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC). July-2019 edn, vol. July-2019, 8981378, Proceedings of 2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC), no. July-2019, IEEE, Xiamen, China, pp. 276-280, 2019 4th IEEE International Conference on Image, Vision and Computing, Xiamen, China, 5/07/19. https://doi.org/10.1109/ICIVC47709.2019.8981378

Towards Content Independent No-reference Image Quality Assessment Using Deep Learning. / Lu, Tan ; Dooms, Ann.
2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC). Vol. July-2019 July-2019. ed. Xiamen, China: IEEE, 2019. p. 276-280 8981378 (Proceedings of 2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC); No. July-2019).

Research output: Chapter in Book/Report/Conference proceeding › Conference paper

TY - GEN

T1 - Towards Content Independent No-reference Image Quality Assessment Using Deep Learning

AU - Lu, Tan

AU - Dooms, Ann

PY - 2019/7

Y1 - 2019/7

N2 - The study of image quality assessment (IQA) is divided on natural scene and document images which are processed using different models and quality metrics. This casts challenges for the development of content-independent no-reference (NR) IQA models which can operate on different types of images without requiring information regarding the content of the images. In this paper we propose a unified no-reference image quality assessment (UIQA) model using a deep learning approach, where a generalization of NR IQA across natural scene and document images is achieved using a deep convolutional neural network (DCNN). Without having to discriminate the type of the images, the proposed model can assess the quality of natural scene and document images in a blind and uniform manner. Testing results on two benchmarking datasets demonstrate that the proposed model achieves promising performances competitive with the state-of-the-art simultaneously on natural scene and document images.

AB - The study of image quality assessment (IQA) is divided on natural scene and document images which are processed using different models and quality metrics. This casts challenges for the development of content-independent no-reference (NR) IQA models which can operate on different types of images without requiring information regarding the content of the images. In this paper we propose a unified no-reference image quality assessment (UIQA) model using a deep learning approach, where a generalization of NR IQA across natural scene and document images is achieved using a deep convolutional neural network (DCNN). Without having to discriminate the type of the images, the proposed model can assess the quality of natural scene and document images in a blind and uniform manner. Testing results on two benchmarking datasets demonstrate that the proposed model achieves promising performances competitive with the state-of-the-art simultaneously on natural scene and document images.

KW - no-reference image quality assessment

KW - perceptual score

KW - OCR accuracy

KW - deep convolutional neural network

KW - transfer learning

UR - https://ieeexplore.ieee.org/document/8981378

UR - https://www.scopus.com/pages/publications/85084293008

U2 - 10.1109/ICIVC47709.2019.8981378

DO - 10.1109/ICIVC47709.2019.8981378

M3 - Conference paper

SN - 978-1-7281-2326-4

VL - July-2019

T3 - Proceedings of 2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC)

SP - 276

EP - 280

BT - 2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC)

PB - IEEE

CY - Xiamen, China

T2 - 2019 4th IEEE International Conference on Image, Vision and Computing

Y2 - 5 July 2019 through 7 July 2019

ER -

Lu T , Dooms A. Towards Content Independent No-reference Image Quality Assessment Using Deep Learning. In 2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC). July-2019 ed. Vol. July-2019. Xiamen, China: IEEE. 2019. p. 276-280. 8981378. (Proceedings of 2019 IEEE 4th International Conference on Image, Vision and Computing (ICIVC); July-2019). doi: 10.1109/ICIVC47709.2019.8981378

Towards Content Independent No-reference Image Quality Assessment Using Deep Learning

Abstract

Publication series

Conference

Keywords

Access to Document

Handle.net

Other files and links

Fingerprint

A Deep Transfer Learning Approach to Document Image Quality Assessment