Reference Quality Assessment

Reference-free quality assessment (QA) aims to objectively evaluate the perceptual quality of multimedia content (images, videos, audio, and 3D models) without needing a pristine reference version. Current research heavily utilizes deep learning, particularly convolutional neural networks (CNNs) and transformers, often incorporating multi-modal features and specialized architectures tailored to specific content types (e.g., handling motion in videos or angular consistency in light fields). These advancements are crucial for improving the efficiency and accuracy of multimedia processing pipelines and enabling fair comparisons of algorithms across diverse applications, from video compression to augmented reality.