HiVE-VQA: Hierarchical Interaction via Volumetric Embedding for Video Quality Assessment
R. Hamilton, F. Zhang, D. Bull
IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2027
A no-reference video quality assessment model built on a JEPA-pretrained encoder with a hierarchical volumetric quality head. Benchmarked against 17 state-of-the-art metrics across 4 datasets, outperforming all of them. Trained with a distributed HPC setup across 8 GPUs.
