Study of Subjective and Objective Quality in Super-Resolution Enhanced Broadcast Images on a Novel SR-IQA Dataset

2024-09-26

arXiv.org

AI Summary

Evaluating the perceptual quality of super-resolved (SR) broadcast images without access to original high-resolution references remains a critical challenge. Method: We construct SR-IQA, the first large-scale subjective dataset for SR image quality assessment, covering 2K/4K resolutions and comprising diverse low-quality broadcast content alongside their SR outputs, accompanied by extensive Mean Opinion Score (MOS) annotations. Through systematic benchmarking and factor analysis, we find that mainstream full-reference and no-reference IQA metrics—including PSNR, LPIPS, and NIQE—exhibit poor correlation with human perception (SROCC < 0.3), exposing fundamental limitations in SR scenarios. Contribution/Results: We introduce the first perception-driven factor system tailored to broadcast SR imagery, demonstrating that existing IQA methods cannot support closed-loop quality optimization. This work establishes a high-fidelity data foundation, empirical evidence, and theoretical framework for developing next-generation perceptually aligned IQA models.

Technology Category

• Computer Vision: Image and Video Retrieval
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Knowledge Representation and Reasoning: Qualitative Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

To display low-quality broadcast content on high-resolution screens in full-screen format, the application of Super-Resolution (SR), a key consumer technology, is essential. Recently, SR methods have been developed that not only increase resolution while preserving the original image information but also enhance the perceived quality. However, evaluating the quality of SR images generated from low-quality sources, such as SR-enhanced broadcast content, is challenging due to the need to consider both distortions and improvements. Additionally, assessing SR image quality without original high-quality sources presents another significant challenge. Unfortunately, there has been a dearth of research specifically addressing the Image Quality Assessment (IQA) of SR images under these conditions. In this work, we introduce a new IQA dataset for SR broadcast images in both 2K and 4K resolutions. We conducted a subjective quality evaluation to obtain the Mean Opinion Score (MOS) for these SR images and performed a comprehensive human study to identify the key factors influencing the perceived quality. Finally, we evaluated the performance of existing IQA metrics on our dataset. This study reveals the limitations of current metrics, highlighting the need for a more robust IQA metric that better correlates with the perceived quality of SR images.

Problem

• Evaluating SR image quality without original high-quality references
• Assessing both distortions and improvements in SR-enhanced broadcast content
• Developing robust IQA metrics correlating with perceived SR image quality

Innovation

• Created novel SR-IQA dataset for broadcast images
• Conducted subjective quality evaluation using MOS
• Evaluated existing IQA metrics on SR images