List your company
ResearchFormerly WizWand

Video Quality Assessment on KoNViD-1k (10 train-test splits)

0.909SROCCDOVER
0.761320.799660.8380.87634Nov 9, 2022
Updated 8mo ago

Evaluation results

MethodLinks
DOVERSource (Pre-training) Quality Dataset=LSVQ [1], Fine-tuning=true2022.110.9090.906
FAST-VQASource (Pre-training) Quality Dataset=LSVQ [1], Fine-tuning=true2022.110.8910.892
Li et al.Source (Pre-training) Quality Dataset=fused ([15,73-75]), Fine-tuning=true2022.110.8340.836
CNN-TLVQMSource (Pre-training) Quality Dataset=handcraft + KoNiQ [73], Fine-tuning=true2022.110.8160.818
CNN-VIDEVALSource (Pre-training) Quality Dataset=handcraft + KoNiQ [73], Fine-tuning=true2022.110.8150.817
RAPIQUESource (Pre-training) Quality Dataset=handcraft + KoNiQ [73], Fine-tuning=true2022.110.8030.817
Patch-VQSource (Pre-training) Quality Dataset=PaQ-2-PiQ [60], Fine-tuning=true2022.110.7910.786
VIDEVALSource (Pre-training) Quality Dataset=NA (pure handcraft), Fine-tuning=true2022.110.7830.78
TLVQMSource (Pre-training) Quality Dataset=NA (pure handcraft), Fine-tuning=true2022.110.7730.768
VSFASource (Pre-training) Quality Dataset=None, Fine-tuning=true2022.110.7730.775
COINVQSource (Pre-training) Quality Dataset=self-collected, Fine-tuning=true2022.110.7670.764