Machine Learning-Based Blind Image Quality Assessment: A Review

In the era of artificial intelligence, the necessity of a paradigm that can characterise the accuracy, fairness, and consequences of decision-making models is highly desirable. The advancement of IoT-enabled multimedia devices and their increasing usage in complex domains such as intelligent remote sensing and autonomous underwater surveillance require extraction of comprehensive patterns from unstructured data. However, the accuracy of such decision-making models solely depends on the quality of data acquisition, transmission, and processing, which gives rise to the necessity of automated assessment of the quality of image data. Hence, this work reviews the state-of-the-art machine learning-based blind image quality assessment algorithms and discusses their possible limitations.