Video Quality Assessment

Video Quality Assessment is a task in the field of computer vision that aims to model human subjective perception based on videos. Its goal is to generate Mean Opinion Score (MOS) values to evaluate the perceptual quality of videos, with higher scores indicating better quality. This task has significant application value in video coding, transmission, and processing. Commonly used benchmark datasets include KoNViD-1k, LIVE-VQC, YouTube-UGC, and LSVQ, and performance evaluation metrics typically adopt SROCC, PLCC, and RMSE.