Perceptual Quality
Perceptual quality assessment focuses on objectively measuring how humans perceive the quality of images and videos, aiming to create models that accurately reflect subjective ratings. Current research emphasizes developing no-reference (NR) methods, particularly using deep learning architectures like transformers and diffusion models, to evaluate diverse content including high-resolution photos, 360° images, and user-generated videos. This field is crucial for improving image and video processing techniques, optimizing compression algorithms, and enhancing the user experience in various applications such as virtual reality and social media.
Papers
A General Method to Incorporate Spatial Information into Loss Functions for GAN-based Super-resolution Models
Xijun Wang, Santiago López-Tapia, Alice Lucas, Xinyi Wu, Rafael Molina, Aggelos K. Katsaggelos
Perceptual Quality-based Model Training under Annotator Label Uncertainty
Chen Zhou, Mohit Prabhushankar, Ghassan AlRegib
PQDynamicISP: Dynamically Controlled Image Signal Processor for Any Image Sensors Pursuing Perceptual Quality
Masakazu Yoshimura, Junji Otsuka, Takeshi Ohashi