Perceptual Quality Assessment of Virtual Reality Videos in the Wild: Insights and Analysis

Core Concepts

Investigating human perception of VR videos in the wild is crucial for quality assessment.

Abstract

The study focuses on understanding how people perceive user-generated VR videos, analyzing viewing behaviors, and developing a blind VQA model. The VRVQW database contains diverse videos with authentic distortions. A psychophysical experiment recorded scanpaths and quality scores under different viewing conditions. Results show significant impact of viewing conditions on perceived quality. A blind VQA model based on pseudocylindrical representation outperforms existing models.

Stats

VRVQW database contains 502 user-generated videos.
139 participants rated video quality under two viewing conditions.
Model performance superior to existing methods.

Quotes

Key Insights Distilled From

Perceptual Quality Assessment of Virtual Reality Videos in the Wild

by Wen Wen,Mu L... at arxiv.org 03-18-2024

https://arxiv.org/pdf/2206.08751.pdf

Deeper Inquiries

How do authentic distortions impact the perceived quality of VR videos

Authentic distortions in VR videos impact the perceived quality by introducing complex visual artifacts that arise during the creation of user-generated 360° videos. These distortions can include under/over-exposure, motion blurring, abrupt luminance or structure changes, objects with missing parts, ghosting, and artificial converging points visible at the poles. They are often localized spatiotemporally and can vary in intensity throughout the video. These authentic distortions can significantly affect how users perceive the content, leading to differences in perceived quality based on viewing conditions such as starting point and exploration time.

Is there a need for standardized subjective procedures in VR video quality assessment

Standardized subjective procedures are essential in VR video quality assessment to ensure consistency and reliability in collecting human perceptual data. By following established protocols like those outlined in ITU-R BT 500.13 recommendation for single stimulus continuous quality evaluation method, researchers can gather MOSs (Mean Opinion Scores) from participants under controlled conditions. Standardization helps minimize biases and ensures that subjective ratings accurately reflect viewers' perceptions of VR video quality across different viewing conditions.

How can the findings from this study be applied to improve user experience in virtual reality applications

The findings from this study have practical implications for improving user experience in virtual reality applications. By understanding how individuals perceive visual distortions in VR environments under varying viewing conditions, developers can optimize content creation processes to minimize negative effects caused by authentic distortions. Additionally, insights into factors influencing perceived quality such as spatial resolution, camera motion, and exploration time can guide enhancements to VR systems and platforms to deliver more immersive and engaging experiences for users.