VideoSet : A large-scale compressed video quality dataset based on JND measurement

Haiqiang WANG; Ioannis KATSAVOUNIDIS; Jiantong ZHOU; Jeonghoon PARK; Shawmin LEI; Xin ZHOU; Man-On PUN; Xin JIN; Ronggang WANG; Xu WANG; Yun ZHANG; Jiwu HUANG; Sam KWONG; C.-C. Jay KUO

doi:10.1016/j.jvcir.2017.04.009

VideoSet : A large-scale compressed video quality dataset based on JND measurement

Haiqiang WANG, Ioannis KATSAVOUNIDIS, Jiantong ZHOU, Jeonghoon PARK, Shawmin LEI, Xin ZHOU, Man-On PUN, Xin JIN, Ronggang WANG, Xu WANG, Yun ZHANG, Jiwu HUANG, Sam KWONG, C.-C. Jay KUO

Research output: Journal Publications › Journal Article (refereed) › peer-review

120 Citations (Scopus)

Abstract

A new methodology to measure coded image/video quality using the just-noticeable-difference (JND) idea was proposed in Lin et al. (2015). Several small JND-based image/video quality datasets were released by the Media Communications Lab at the University of Southern California in Jin et al. (2016) and Wang et al. (2016) [3]. In this work, we present an effort to build a large-scale JND-based coded video quality dataset. The dataset consists of 220 5-s sequences in four resolutions (i.e., 1920×1080,1280×720,960×540 and 640×360). For each of the 880 video clips, we encode it using the H.264/AVC codec with QP=1,…,51 and measure the first three JND points with 30 + subjects. The dataset is called the “VideoSet”, which is an acronym for “Video Subject Evaluation Test (SET)”. This work describes the subjective test procedure, detection and removal of outlying measured data, and the properties of collected JND data. Finally, the significance and implications of the VideoSet to future video coding research and standardization efforts are pointed out. All source/coded video clips as well as measured JND data included in the VideoSet are available to the public in the IEEE DataPort (Wang et al., 2016 [4]).

Original language	English
Pages (from-to)	292-302
Journal	Journal of Visual Communication and Image Representation
Volume	46
DOIs	https://doi.org/10.1016/j.jvcir.2017.04.009
Publication status	Published - 1 Jul 2017
Externally published	Yes

Bibliographical note

This research was funded by Netflix, Huawei, Samsung and MediaTek. The subjective tests were conducted in the City University of Hong Kong and five universities in the Shenzhen City of China. They were Shenzhen University, Chinese University of Hong Kong (SZ), Tsinghua University, Peking University and Chinese Academy of Sciences. Computation for the work was supported in part by the University of Southern California's Center for High-Performance Computing (hpc.usc.edu). The authors would like to give thanks to these companies and universities for their strong support.

Keywords

Human visual system (HVS)
Just noticeable difference (JND)
Video coding
Video quality

Access to Document

10.1016/j.jvcir.2017.04.009

Cite this

@article{1b971f25b8ba4b77a16a22a83cb4665e,

title = "VideoSet : A large-scale compressed video quality dataset based on JND measurement",

abstract = "A new methodology to measure coded image/video quality using the just-noticeable-difference (JND) idea was proposed in Lin et al. (2015). Several small JND-based image/video quality datasets were released by the Media Communications Lab at the University of Southern California in Jin et al. (2016) and Wang et al. (2016) [3]. In this work, we present an effort to build a large-scale JND-based coded video quality dataset. The dataset consists of 220 5-s sequences in four resolutions (i.e., 1920×1080,1280×720,960×540 and 640×360). For each of the 880 video clips, we encode it using the H.264/AVC codec with QP=1,…,51 and measure the first three JND points with 30 + subjects. The dataset is called the “VideoSet”, which is an acronym for “Video Subject Evaluation Test (SET)”. This work describes the subjective test procedure, detection and removal of outlying measured data, and the properties of collected JND data. Finally, the significance and implications of the VideoSet to future video coding research and standardization efforts are pointed out. All source/coded video clips as well as measured JND data included in the VideoSet are available to the public in the IEEE DataPort (Wang et al., 2016 [4]).",

keywords = "Human visual system (HVS), Just noticeable difference (JND), Video coding, Video quality",

author = "Haiqiang WANG and Ioannis KATSAVOUNIDIS and Jiantong ZHOU and Jeonghoon PARK and Shawmin LEI and Xin ZHOU and Man-On PUN and Xin JIN and Ronggang WANG and Xu WANG and Yun ZHANG and Jiwu HUANG and Sam KWONG and KUO, {C.-C. Jay}",

note = "This research was funded by Netflix, Huawei, Samsung and MediaTek. The subjective tests were conducted in the City University of Hong Kong and five universities in the Shenzhen City of China. They were Shenzhen University, Chinese University of Hong Kong (SZ), Tsinghua University, Peking University and Chinese Academy of Sciences. Computation for the work was supported in part by the University of Southern California's Center for High-Performance Computing (hpc.usc.edu). The authors would like to give thanks to these companies and universities for their strong support.",

year = "2017",

month = jul,

day = "1",

doi = "10.1016/j.jvcir.2017.04.009",

language = "English",

volume = "46",

pages = "292--302",

journal = "Journal of Visual Communication and Image Representation",

issn = "1047-3203",

publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - VideoSet : A large-scale compressed video quality dataset based on JND measurement

AU - WANG, Haiqiang

AU - KATSAVOUNIDIS, Ioannis

AU - ZHOU, Jiantong

AU - PARK, Jeonghoon

AU - LEI, Shawmin

AU - ZHOU, Xin

AU - PUN, Man-On

AU - JIN, Xin

AU - WANG, Ronggang

AU - WANG, Xu

AU - ZHANG, Yun

AU - HUANG, Jiwu

AU - KWONG, Sam

AU - KUO, C.-C. Jay

N1 - This research was funded by Netflix, Huawei, Samsung and MediaTek. The subjective tests were conducted in the City University of Hong Kong and five universities in the Shenzhen City of China. They were Shenzhen University, Chinese University of Hong Kong (SZ), Tsinghua University, Peking University and Chinese Academy of Sciences. Computation for the work was supported in part by the University of Southern California's Center for High-Performance Computing (hpc.usc.edu). The authors would like to give thanks to these companies and universities for their strong support.

PY - 2017/7/1

Y1 - 2017/7/1

N2 - A new methodology to measure coded image/video quality using the just-noticeable-difference (JND) idea was proposed in Lin et al. (2015). Several small JND-based image/video quality datasets were released by the Media Communications Lab at the University of Southern California in Jin et al. (2016) and Wang et al. (2016) [3]. In this work, we present an effort to build a large-scale JND-based coded video quality dataset. The dataset consists of 220 5-s sequences in four resolutions (i.e., 1920×1080,1280×720,960×540 and 640×360). For each of the 880 video clips, we encode it using the H.264/AVC codec with QP=1,…,51 and measure the first three JND points with 30 + subjects. The dataset is called the “VideoSet”, which is an acronym for “Video Subject Evaluation Test (SET)”. This work describes the subjective test procedure, detection and removal of outlying measured data, and the properties of collected JND data. Finally, the significance and implications of the VideoSet to future video coding research and standardization efforts are pointed out. All source/coded video clips as well as measured JND data included in the VideoSet are available to the public in the IEEE DataPort (Wang et al., 2016 [4]).

AB - A new methodology to measure coded image/video quality using the just-noticeable-difference (JND) idea was proposed in Lin et al. (2015). Several small JND-based image/video quality datasets were released by the Media Communications Lab at the University of Southern California in Jin et al. (2016) and Wang et al. (2016) [3]. In this work, we present an effort to build a large-scale JND-based coded video quality dataset. The dataset consists of 220 5-s sequences in four resolutions (i.e., 1920×1080,1280×720,960×540 and 640×360). For each of the 880 video clips, we encode it using the H.264/AVC codec with QP=1,…,51 and measure the first three JND points with 30 + subjects. The dataset is called the “VideoSet”, which is an acronym for “Video Subject Evaluation Test (SET)”. This work describes the subjective test procedure, detection and removal of outlying measured data, and the properties of collected JND data. Finally, the significance and implications of the VideoSet to future video coding research and standardization efforts are pointed out. All source/coded video clips as well as measured JND data included in the VideoSet are available to the public in the IEEE DataPort (Wang et al., 2016 [4]).

KW - Human visual system (HVS)

KW - Just noticeable difference (JND)

KW - Video coding

KW - Video quality

UR - http://www.scopus.com/inward/record.url?scp=85019029618&partnerID=8YFLogxK

U2 - 10.1016/j.jvcir.2017.04.009

DO - 10.1016/j.jvcir.2017.04.009

M3 - Journal Article (refereed)

SN - 1047-3203

VL - 46

SP - 292

EP - 302

JO - Journal of Visual Communication and Image Representation

JF - Journal of Visual Communication and Image Representation

ER -

VideoSet : A large-scale compressed video quality dataset based on JND measurement

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this