Merging of MOS of Large Image Databases for No-reference Image Visual Quality Assessment

Aki Kaipio; Mykola Ponomarenko; Karen Egiazarian

doi:10.1109/MMSP48831.2020.9287141

Merging of MOS of Large Image Databases for No-reference Image Visual Quality Assessment

Aki Kaipio, Mykola Ponomarenko, Karen Egiazarian

Computing Sciences

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Scientific › peer-review

6 Citations (Scopus)

Abstract

For training of no-reference image visual quality metrics large specialized image databases are used. For images of the databases mean opinion scores (MOS) are experimentally obtained collecting judgments of many observers. MOS of a given image reflects an averaged human perception of visual quality of the image. Each database has its own unknown scale of MOS values depending on unique content of the database. For training of no-reference metrics based on convolutional networks usually only one selected database is used, because all MOS values on input of training loss function should be in the same scale. In this paper, a simple and effective method of merging of several large databases into one database with transforming of their MOS into one scale is proposed. Accuracy of the proposed method is analyzed. Merged MOS is used for practical training of no-reference metric. Better effectiveness of the training is shown in comparative analysis.

Original language	English
Title of host publication	2020 IEEE 22nd International Workshop on Multimedia Signal Processing (MMSP)
Publisher	IEEE
Number of pages	6
ISBN (Electronic)	978-1-7281-9320-5
DOIs	https://doi.org/10.1109/MMSP48831.2020.9287141
Publication status	Published - 1 Sept 2020
Publication type	A4 Article in conference proceedings
Event	IEEE International Workshop on Multimedia Signal Processing - Tampere, Finland Duration: 21 Sept 2020 → 24 Sept 2020 https://attend.ieee.org/mmsp-2020/

Publication series

Name	IEEE International Workshop on Multimedia Signal Processing
ISSN (Electronic)	2473-3628

Conference

Conference	IEEE International Workshop on Multimedia Signal Processing
Country/Territory	Finland
City	Tampere
Period	21/09/20 → 24/09/20
Internet address	https://attend.ieee.org/mmsp-2020/

Keywords

Training
Measurement
Visualization
Databases
Image databases
Merging
Visual databases
no-reference image visual quality assessment
deep convolutional networks
image databases with MOS

Publication forum classification

Publication forum level 1

Access to Document

10.1109/MMSP48831.2020.9287141

Cite this

@inproceedings{1e97e022fa5e4af7815b1af46efc6151,

title = "Merging of MOS of Large Image Databases for No-reference Image Visual Quality Assessment",

abstract = "For training of no-reference image visual quality metrics large specialized image databases are used. For images of the databases mean opinion scores (MOS) are experimentally obtained collecting judgments of many observers. MOS of a given image reflects an averaged human perception of visual quality of the image. Each database has its own unknown scale of MOS values depending on unique content of the database. For training of no-reference metrics based on convolutional networks usually only one selected database is used, because all MOS values on input of training loss function should be in the same scale. In this paper, a simple and effective method of merging of several large databases into one database with transforming of their MOS into one scale is proposed. Accuracy of the proposed method is analyzed. Merged MOS is used for practical training of no-reference metric. Better effectiveness of the training is shown in comparative analysis.",

keywords = "Training, Measurement, Visualization, Databases, Image databases, Merging, Visual databases, no-reference image visual quality assessment, deep convolutional networks, image databases with MOS",

author = "Aki Kaipio and Mykola Ponomarenko and Karen Egiazarian",

note = "jufoid=70574; IEEE International Workshop on Multimedia Signal Processing ; Conference date: 21-09-2020 Through 24-09-2020",

year = "2020",

month = sep,

day = "1",

doi = "10.1109/MMSP48831.2020.9287141",

language = "English",

series = "IEEE International Workshop on Multimedia Signal Processing",

publisher = "IEEE",

booktitle = "2020 IEEE 22nd International Workshop on Multimedia Signal Processing (MMSP)",

url = "https://attend.ieee.org/mmsp-2020/",

}

Kaipio, A, Ponomarenko, M & Egiazarian, K 2020, Merging of MOS of Large Image Databases for No-reference Image Visual Quality Assessment. in 2020 IEEE 22nd International Workshop on Multimedia Signal Processing (MMSP). IEEE International Workshop on Multimedia Signal Processing, IEEE, IEEE International Workshop on Multimedia Signal Processing, Tampere, Finland, 21/09/20. https://doi.org/10.1109/MMSP48831.2020.9287141

Merging of MOS of Large Image Databases for No-reference Image Visual Quality Assessment. / Kaipio, Aki; Ponomarenko, Mykola ; Egiazarian, Karen.
2020 IEEE 22nd International Workshop on Multimedia Signal Processing (MMSP). IEEE, 2020. (IEEE International Workshop on Multimedia Signal Processing).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Merging of MOS of Large Image Databases for No-reference Image Visual Quality Assessment

AU - Kaipio, Aki

AU - Ponomarenko, Mykola

AU - Egiazarian, Karen

N1 - jufoid=70574

PY - 2020/9/1

Y1 - 2020/9/1

N2 - For training of no-reference image visual quality metrics large specialized image databases are used. For images of the databases mean opinion scores (MOS) are experimentally obtained collecting judgments of many observers. MOS of a given image reflects an averaged human perception of visual quality of the image. Each database has its own unknown scale of MOS values depending on unique content of the database. For training of no-reference metrics based on convolutional networks usually only one selected database is used, because all MOS values on input of training loss function should be in the same scale. In this paper, a simple and effective method of merging of several large databases into one database with transforming of their MOS into one scale is proposed. Accuracy of the proposed method is analyzed. Merged MOS is used for practical training of no-reference metric. Better effectiveness of the training is shown in comparative analysis.

AB - For training of no-reference image visual quality metrics large specialized image databases are used. For images of the databases mean opinion scores (MOS) are experimentally obtained collecting judgments of many observers. MOS of a given image reflects an averaged human perception of visual quality of the image. Each database has its own unknown scale of MOS values depending on unique content of the database. For training of no-reference metrics based on convolutional networks usually only one selected database is used, because all MOS values on input of training loss function should be in the same scale. In this paper, a simple and effective method of merging of several large databases into one database with transforming of their MOS into one scale is proposed. Accuracy of the proposed method is analyzed. Merged MOS is used for practical training of no-reference metric. Better effectiveness of the training is shown in comparative analysis.

KW - Training

KW - Measurement

KW - Visualization

KW - Databases

KW - Image databases

KW - Merging

KW - Visual databases

KW - no-reference image visual quality assessment

KW - deep convolutional networks

KW - image databases with MOS

U2 - 10.1109/MMSP48831.2020.9287141

DO - 10.1109/MMSP48831.2020.9287141

M3 - Conference contribution

T3 - IEEE International Workshop on Multimedia Signal Processing

BT - 2020 IEEE 22nd International Workshop on Multimedia Signal Processing (MMSP)

PB - IEEE

T2 - IEEE International Workshop on Multimedia Signal Processing

Y2 - 21 September 2020 through 24 September 2020

ER -

Merging of MOS of Large Image Databases for No-reference Image Visual Quality Assessment

Abstract

Publication series

Conference

Keywords

Publication forum classification

Access to Document

Fingerprint

Cite this