Blind Image Quality Assessment (BIQA) is essential for automatically
evaluating the perceptual quality of visual signals without access to the
references. In this survey, we provide a comprehensive analysis and discussion
of recent developments in the field of BIQA. We have covered various aspects,
including hand-crafted BIQAs that focus on distortion-specific and
general-purpose methods, as well as deep-learned BIQAs that employ supervised
and unsupervised learning techniques. Additionally, we have explored multimodal
quality assessment methods that consider interactions between visual and audio
modalities, as well as visual and text modalities. Finally, we have offered
insights into representative BIQA databases, including both synthetic and
authentic distortions. We believe this survey provides valuable understandings
into the latest developments and emerging trends for the visual quality
community.
Blind Image Quality Assessment: A Comprehensive Analysis and Discussion
Blind Image Quality Assessment (BIQA) plays a crucial role in evaluating the perceptual quality of visual signals without the need for reference images. In this survey, we delve into the various developments in the field of BIQA, providing a thorough analysis and discussion.
Hand-crafted BIQAs: Distortion-specific and General-purpose Methods
One of the key aspects of BIQA is the utilization of hand-crafted methods. These approaches are designed to specifically target certain types of distortions, such as noise, blur, or compression artifacts. By focusing on distortion-specific methods, researchers can develop algorithms with a deep understanding of the underlying artifacts and their impact on image quality.
On the other hand, general-purpose BIQAs aim to assess overall image quality by considering a range of potential distortions. These methods take into account a variety of visual features and statistical measures to estimate the quality of an image. By adopting a more holistic approach, general-purpose methods can provide a comprehensive evaluation of visual signals.
Deep-learned BIQAs: Supervised and Unsupervised Learning Techniques
In recent years, deep learning has emerged as a powerful tool in various fields, including BIQA. Deep-learned BIQAs leverage the capabilities of neural networks to learn complex relationships between image features and perceived quality. These approaches can be categorized into supervised and unsupervised learning techniques.
Supervised learning techniques train the neural networks using large datasets that have been annotated with subjective quality scores. This enables the network to learn from human judgments and make accurate predictions about the quality of unseen images. Unsupervised learning techniques, on the other hand, aim to discover inherent structures and patterns within the data without relying on explicit quality labels.
Multimodal Quality Assessment: Visual, Audio, and Text Modalities
One of the intriguing aspects of BIQA is the exploration of multimodal quality assessment methods. These approaches consider the interactions between different modalities, such as visual, audio, and text, to determine overall quality. By incorporating multiple modalities, researchers can capture a more comprehensive understanding of visual signals and their perceived quality.
Representative BIQA Databases: Synthetic and Authentic Distortions
The availability of high-quality databases is crucial for the advancement of BIQA research. This survey highlights the importance of representative BIQA databases that encompass both synthetic and authentic distortions. Synthetic distortions allow researchers to create controlled environments for testing and evaluating algorithms, while authentic distortions reflect real-world scenarios and challenges.
The Wider Field: Multimedia Information Systems, Animations, Artificial Reality, Augmented Reality, and Virtual Realities
The concepts and developments discussed in this survey have strong connections to the wider field of multimedia information systems. Multimedia information systems deal with the storage, retrieval, and analysis of multimedia data, which includes images, videos, animations, and more.
Moreover, the advancements in BIQA impact various applications that fall under the umbrella of artificial reality. Animations, artificial reality, augmented reality, and virtual realities heavily rely on high-quality visual signals to create immersive and realistic experiences. The ability to automatically assess the quality of these visual signals contributes to enhancing the overall user experience.
In conclusion, this survey provides valuable insights into the latest developments and emerging trends in Blind Image Quality Assessment. By covering various approaches, modalities, and databases, it offers a comprehensive understanding of this multi-disciplinary field. As multimedia information systems continue to evolve and intersect with the realms of artificial and virtual reality, BIQA remains an integral component in ensuring high-quality visual content.