Naslov
Comparison of objective video quality assessment methods on videos with different spatial resolutions

Autori
Vlaović, Jelena ; Vranješ, Mario ; Grabić, Dominik ; Samardžija, Dragan

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Proceedings of IWSSIP 2019 / Rima-Drlje, Snježana ; Žagar, Drago ; Galić, Irena ; Martinović, Goran ; Vranješ, Denis ; Habijan, Marija - Osijek : FERIT Osijek, 2019, 287-292

ISBN
978-1-7281-3253-2

Skup
26th International Conference on Systems, Signals and Image Processing (IWSSIP 2019)

Mjesto i datum
Osijek, Hrvatska, 05.06.2019. - 07.06.2019

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
H.264 ; PSNR ; SSIM ; MSSIM ; VQM ; VSNR, spatial and temporal activity

Sažetak
A significant progress has been made in adaptive video streaming technology in recent years, but there is still room for improvement using different optimization methods. The goal is to ensure the highest possible video quality perceived by the end users, thus ensuring the highest possible user’s quality of experience (QoE), regardless to the network conditions. Apart from different coding methods, the content itself and the user experience should be taken into account as optimization parameters. This paper provides the comparison of five objective video quality assessment (VQA) metrics using different video sequences encoded according to H.264/AVC norm at different spatial resolutions and with different target coding bitrates. The metrics included in comparison are Peak Signal-to-Noise Ratio (PSNR), Video Quality Metric (VQM), Structural Similarity index (SSIM), Mean Structural Similarity (MSSIM) index, and Visual Signal-to-noise Ratio (VSNR). Results of objective VQA metrics have been compared to those of performed subjective VQA experiments. The results show that SSIM and MSSIM achieve the highest performance when evaluating the quality of video sequences of different contents, resolutions and coded with a wide range of coding parameters. Additionally, the results indicate that it is more difficult to efficiently compress the video sequences with higher spatial and temporal activity. Also, due to spatial and temporal masking of compression artifacts in high complexity videos, the analyzed objective VQA metrics achieve lower quality prediction accuracy for that videos.

Izvorni jezik
Engleski

Znanstvena područja
Elektrotehnika, Računarstvo

POVEZANOST RADA