Subjective and Objective Quality Assessment of Image: A Survey (1406.7799v1)

Abstract: With the increasing demand for image-based applications, the efficient and reliable evaluation of image quality has increased in importance. Measuring the image quality is of fundamental importance for numerous image processing applications, where the goal of image quality assessment (IQA) methods is to automatically evaluate the quality of images in agreement with human quality judgments. Numerous IQA methods have been proposed over the past years to fulfill this goal. In this paper, a survey of the quality assessment methods for conventional image signals, as well as the newly emerged ones, which includes the high dynamic range (HDR) and 3-D images, is presented. A comprehensive explanation of the subjective and objective IQA and their classification is provided. Six widely used subjective quality datasets, and performance measures are reviewed. Emphasis is given to the full-reference image quality assessment (FR-IQA) methods, and 9 often-used quality measures (including mean squared error (MSE), structural similarity index (SSIM), multi-scale structural similarity index (MS-SSIM), visual information fidelity (VIF), most apparent distortion (MAD), feature similarity measure (FSIM), feature similarity measure for color images (FSIMC), dynamic range independent measure (DRIM), and tone-mapped images quality index (TMQI)) are carefully described, and their performance and computation time on four subjective quality datasets are evaluated. Furthermore, a brief introduction to 3-D IQA is provided and the issues related to this area of research are reviewed.

Image Quality Assessment: A Comprehensive Survey

The paper "Subjective and Objective Quality Assessment of Image: A Survey" provides an exhaustive overview of image quality assessment (IQA) methods focused on aligning with human quality judgments. The authors, Pedram Mohammadi, Abbas Ebrahimi-Moghadam, and Shahram Shirani, categorize IQA into subjective and objective mechanisms, emphasizing the transition toward objective methods given the impracticality of subjective evaluations for real-world applications.

Overview of Image Quality Assessment

Image quality assessment (IQA) has gained prominence due to the ubiquity of image-based applications and the inherent need for reliable quality evaluation mechanisms. Many image processing systems necessitate quality control to manage visual communication and inform decisions concerning image acquisition, compression, and transmission. The paper delineates the pivotal role of IQA across various domains, including biomedical imaging, printing, and modern imaging technologies like high dynamic range (HDR) and 3-D imaging.

Subjective IQA

The paper elucidates subjective IQA methods, which are deemed the most accurate due to their reliance on human observers. However, these evaluations are constrained by cost, time, viewing conditions, and environmental factors affecting participants, leading to inconsistency. Various international methodologies, such as ITU standards, provide guidance on conducting these assessments, but their limitations necessitate the shift towards objective IQA models, which can simulate human evaluation processes through mathematical approaches.

Objective IQA

Objective IQA methods aim to predict image quality as perceived by humans, using models that do not depend on human intervention. These are classified into:

• Full-reference IQA (FR-IQA): Utilizes a distortion-free reference image for comparisons.
• Reduced-reference IQA (RR-IQA): Employs partial data from reference images due to its limited availability.
• No-reference IQA (NR-IQA): Operates without reference images, relying on test images alone.

The survey emphasizes FR-IQA methods, detailing algorithms covering metrics like SSIM, MS-SSIM, VIF, MAD, FSIM, and FSIMc. Each metric is dissected, exploring their computational strategies, strengths, and typical applications, particularly targeting gray-scale images, color images, and different dynamic ranges inherent in HDR imaging.

Performance and Evaluation

The paper evaluates these algorithms using datasets, assessing correlation measures like PLCC, SRCC, and KRCC to determine prediction efficacy. It also benchmarks computation times, providing practicality insights across varied application environments.

3-D Image Quality Assessment

3-D IQA constitutes a nuanced domain, addressing the complexity beyond traditional 2-D images. The survey touches upon 3-D image datasets and emphasizes the unique considerations for depth perception and stereoscopic distortions. Despite 2-D methods serving symmetric 3-D images effectively, factors unique to 3-D visions, such as Cyclopean images, introduce challenges for traditional assessment models.

Implications and Future Directions

The implications of this survey highlight the need for continued advancements in objective IQA methodologies, driven by computational enhancements and a deeper understanding of human vision and perception correlates. As multimedia applications become more sophisticated, further research into real-time and perceptually aligned IQA models remains imperative.

The paper facilitates future endeavors in developing robust, scalable IQA models that bridge the perceptual gap between computational predictions and human experience, critical for applications ranging from consumer electronics to critical systems in medical and remote sensing fields. As the landscape evolves, embracing novel image types and complex visual inputs will guide comprehensive improvements in image quality assessment technology.