Perceptual reliability in multimedia: Quality assessment and anomaly analysis

Zhou, Wei, Liu, Yang, Amirpour, Hadi and Callet, Patrick Le 2026. Perceptual reliability in multimedia: Quality assessment and anomaly analysis. Presented at: 34th ACM International Conference on Multimedia, Rio de Janeiro, Brazil, 10-14 Nov 2026. Proceedings of the 34th ACM International Conference on Multimedia. ACM International Conference on Multimedia (ACM MM). ACM, pp. 17934-17935. 10.1145/3767308.3836610

PDF - Published Version Available under License Creative Commons Attribution. Download (837kB)

Abstract

The rapid expansion of multimedia services, including video streaming, conferencing, immersive media, and AI-generated content, makes perceptual reliability a central requirement for modern multimedia systems. Multimedia signals can be degraded during acquisition, compression, transmission, enhancement, and rendering, while learning-based components can introduce semantic inconsistencies, model bias, and uncertain predictions. This tutorial presents quality assessment and anomaly detection as two complementary aspects of perceptual reliability. Quality assessment characterizes perceived quality under distortions and content variations, whereas anomaly detection identifies unexpected semantic or behavioral patterns. We connect multimedia representations and distortion formation with human perception, subjective and objective quality assessment, statistical and deep anomaly detection, robustness under domain shift, and multimodal foundation models. We further discuss immersive and AI-generated media, where reliability encompasses spatial comfort, semantic fidelity, realism, temporal and cross-modal consistency, and uncertainty. The tutorial provides a unified, practice-oriented framework for evaluating and improving trustworthy multimedia pipelines.

Item Type:	Conference or Workshop Item - published (Paper)
Date Type:	Publication
Status:	Published
Schools:	Schools > Computational & Mathematical Sciences Schools > Computer Science & Informatics
Publisher:	ACM
ISBN:	9798400722134
Date of First Compliant Deposit:	12 November 2026
Last Modified:	12 Nov 2026 11:30
URI:	https://orca.cardiff.ac.uk/id/eprint/190884

Actions (repository staff only)

Edit Item