2013 IEEE 15th International Workshop on Multimedia Signal Processing (MMSP) 2013
DOI: 10.1109/mmsp.2013.6659319
|View full text |Cite
|
Sign up to set email alerts
|

About the imperfection of objective quality metrics on high-definition video content

Abstract: Due to the ever-growing importance of video content delivered through today's Internet, not only the availability of high-definition video material but also its high quality must be guaranteed. The latter can be achieved by means of objective video quality metrics. Recently, insufficient quality prediction accuracy for well-known full-reference objective quality metrics on high-definition video content has been reported. With this paper we aim at identifying possible reasons for the observed imperfection.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1

Citation Types

0
2
0
1

Year Published

2014
2014
2018
2018

Publication Types

Select...
2
1

Relationship

0
3

Authors

Journals

citations
Cited by 3 publications
(3 citation statements)
references
References 18 publications
(32 reference statements)
0
2
0
1
Order By: Relevance
“…Observing the quantile plot (figures on the right) for the three metrics, the PSNR model is the least accurate of the three because PSNR samples are deviated from the linear distribution. Nonetheless, PSNR has been criticized throughout its history for not serving well as a QoE metric [16] [17]. Yet, the exponential model is not entirely inaccurate, as the observation of residuals reveals.…”
Section: Subjective Results and Objective Metrics Correlationmentioning
confidence: 99%
“…Observing the quantile plot (figures on the right) for the three metrics, the PSNR model is the least accurate of the three because PSNR samples are deviated from the linear distribution. Nonetheless, PSNR has been criticized throughout its history for not serving well as a QoE metric [16] [17]. Yet, the exponential model is not entirely inaccurate, as the observation of residuals reveals.…”
Section: Subjective Results and Objective Metrics Correlationmentioning
confidence: 99%
“…The second inexactness deals with the requirement for temporal and color alignment of the reference and test sequence which was not present in the first versions of PSNR and may be considered as non-normative preprocessing steps [12]. However, the alignment has an important impact on the final result: When temporal mismatch occurs, PSNR without temporal alignment underestimates the quality because it uses the wrong reference for calculation [13]. PSNR with temporal alignment often overestimates the quality as effects such as stalling, skipping, or reduced frame rate are ignored.…”
Section: Introductionmentioning
confidence: 99%
“…De estas métricas, VQM_VFD fue la que mejor resultado consiguió, obteniendo una correlación en torno a 0,81, pese a no estar entrenada específicamente para este tipo de esquema de codificación. En [Wulf and Zolzer, 2013] también se pone de manifiesto las ventajas que supone el nuevo método de calibración VFD en comparación con otras métricas de calidad de vídeo, aplicadas a varias bases de datos de secuencias de vídeo.…”
Section: Métricas De Referencia Completaunclassified