Perceptual video quality metric for compression artefacts: from two‐dimensional to omnidirectional

A Survey on 360-Degree Video: Coding, Quality of Experience and Streaming

“…Another work [138] compares the basic metrics' performance and complexity, and finds that most are well-correlated to MOS in the studied scenarios. The experiments by the authors show that the more complex methods in [130,131,109,132,133] have a higher performance than traditional metrics, but they have not been corroborated by independent studies yet.…”

“…More complex metrics take into account several factors, often combining the objective metrics mentioned above: in [130], a non-linear Perceptual Video Quality (PVQ) model is derived, starting from SSIM and other metrics and matching them to a predicted MOS. The same operation is performed by the Normalized Quality versus Quality factor (NQQ) model in [131], which computes QoE as a non-linear function of a combination of coding parameters such as spatial resolution and quantization factor, whose parameters are derived from the spatial activity in the image and the low-order moments of the luminance distribution.…”

“…Metrics such as the Natural Image Quality Evaluator (NIQE) [140], based on natural image statistics, and the Six-Step Blind Metric (SISBLIM) [141], which is the combination of six different distortion measurements, have good performance on 2D images and videos, but the only study to check their effectiveness for immersive video [111] has found that their performance is significantly affected by the geometric distortion, making them only weakly correlated with subjectively perceived quality. The Multi S-SSIM [127] SSIM with corrections for projective distortion in the spherical domain [127] WS-SSIM [128] SSIM weighted proportionally to pixel area on the sphere [128] Voronoi [114] SSIM and PSNR on Voronoi patches [114] CP-PSNR [99] Saliency-and viewport-weighted PSNR [99] CP-SSIM [99] Saliency-and viewport-weighted SSIM [99] PVQ [130] Non-linear function of SSIM [130] NQQ [131] Non-linear function of the coding parameters [131] QAVR [132] Learning-based model based on features at multiple scales [132] DeepVR-IQA [133] Adversarial generative model to learn QoE [133] Model in [109] Learning-based metric with head and eye movement input [109] V-CNN [134] CNN on viewports weighted by viewing probability [134] Channel 360 • Image Quality Assessment (MC360IQA) metric [142] is a no reference metric using a multi-channel CNN on the six faces of a cube, trained on the dataset in [111]: the metric outperforms even 2D full reference metrics on the dataset.…”

Network video quality assessment method using fuzzy decision tree

“…From entertainment on the move‐hand‐held phones spewing out videos – to entertainment at home, digital videos are everywhere. Moreover, wireless systems are rapidly replacing wire‐line systems, and new‐generation encoders with tremendously improved compression efficiency are being standardised [1, 2]. However, the network video should be compressed and transmitted from the server to the receiver.…”