With the deployment of low-cost sensors, social media platforms, and cloud storage, a tremendous amount of image and video technologies are becoming available cheaply. As a standard tool to analyze these types of data, quality models have been used pervasively in domains including intelligent systems and 3D rendering. Quality models are designed to mimic human visual perception and its underlying mechanism. For example, human-eye shifting paths can be mapped and leveraged for visual quality assessment. Moreover, the function of the human visual cortex can be mathematically modeled and a multi-layer feature fusion algorithm can be proposed to seamlessly combine multiple visual factors and mimic the hierarchy of human visual quality perception. The mechanics and functionality of human and animal visual systems when visual information with different qualities are perceived can also be investigated.

In recent decades, many shallow quality models have been developed and commercialized. Despite their success, conventional quality models may be deficient in the handling of the increasing amount of media data. Owing to the significant progress in deep feature engineering, deep quality models have been proposed with satisfactory performance. For future media platforms, the quality model needs to be human interpretable. However, the existing deep models are conducted in a black-box manner and the means to make visual quality modeling interpretable and transparent with efficient learning are still unresolved. When modeling existing-scale data for future visual systems, compared to fully or partially annotated signals it might be infeasible to label large-scale images and videos at the pixel level given the manual work that would be required to support this. In the future, only image and video-level labels or partial labels may be available however, sometimes such labels can be contaminated. Thus, how to design a noise-tolerant and weakly-supervised learning algorithm for exploring quality-aware features is a difficult problem to address. Conventional quality models typically leverage local or global features to evaluate each image or video, where human visual perception cannot be encoded explicitly. It has been reported that human visual perception plays a significant role in media quality modeling, especially for next-generation visual systems where human-computer interaction and personal profiles are emphasized. Current literature highlights the difficulties in mimicking human visual perception such as predicting human gaze behavior and subsequently modeling the visual signal cognition in the human brain.

This Special Issue focuses on recent progress in image and video quality modeling and analytics. We aim to explore interpretable, noise-tolerant, and biologically-inspired deep models to enhance visual quality models. Submissions focusing on the new image and video benchmarks for testing the performance of quality models are also welcomed. We also welcome submissions investigating the complicated biological mechanism of visual quality perception in humans and animals. The primary objective of this Special Issue is to focus attention on the latest research progress in this cutting-edge area. We welcome submissions of both original research and review articles from researchers and practitioners from both industry and academia.

Potential topics include but are not limited to the following:

• Next-generation deep architectures for image/video/audio quality evaluation
• Deep algorithms for enhancing the shallow-feature-based intelligent systems
• Quality-driven audio/image/video processing techniques
• Visual quality prediction for future photo/video management systems
• Leveraging human interactions to improve deep quality models
• Perception-aware quality models for next-generation scale media retrieval
• Novel deep quality features and their applications in pattern recognition
• Deep models trained using small samples for quality understanding
• Novel photo or video retargeting/cropping/re-composition using deep features
• New datasets, benchmarks, and validation of deep quality models
• Subjective methodologies to estimate the quality in future media systems
• Novel visualization technologies for deep quality features
• Potential challenges of media quality modeling in future intelligent systems