List your company
ResearchFormerly WizWand

Video Quality Assessment on LIVE-VQC (test)

0.849SRCCFAST-VQA
0.475640.572570.66950.76643May 29, 2020Jan 9, 2021Aug 23, 2021Apr 5, 2022Nov 17, 2022Jun 30, 2023Feb 11, 2024
Updated 8mo ago

Evaluation results

MethodLinks
FAST-VQAGroup=Ours, Protocol=Finetune2022.070.8490.865
FAST-VQAResolution Range=240P - 1080P2022.100.8490.865
FasterVQAResolution Range=240P - 1080P, Efficiency=4X than FAST-VQA2022.100.8430.858
DOVERTraining Set=LSVQ_train, GFLOPS=282.3, CPU Time=9.7s, GPU Time=47ms2022.110.8320.855
BVQA-TCSVT-2022Group=Existing Fixed Deep, Resolution Range=240P - 1080P2022.100.8310.842
FAST-VQA-MGroup=Ours, Protocol=Finetune2022.070.8280.849
PVQGroup=Existing Fixed Deep, Protocol=Finetune2022.070.8270.837
PVQGroup=Existing Fixed Deep, Resolution Range=240P - 1080P2022.100.8270.837
CNN+TLVQMGroup=Ensemble C+D, Protocol=Finetune2022.070.8250.834
CNN+TLVQMGroup=Ensemble C+D, Resolution Range=240P - 1080P2022.100.8250.834
FAST-VQATraining Set=LSVQ_train, GFLOPS=279.1, CPU Time=8.8s, GPU Time=45ms2022.110.8230.844
PVQTraining Setting=Mixed Databases2021.080.81620.8445
ProposedTraining Dataset=LSVQ, Evaluation Protocol=Cross-database2021.080.8160.824
Li et al.Training Set=LSVQ_train, GFLOPS=112537, CPU Time=1567s, GPU Time=27.6s2022.110.8160.824
FAST-VQA-MResolution Range=240P - 1080P2022.100.8030.828
TLVQMGroup=Existing Classical, Protocol=Finetune2022.070.7990.803
Full-res Swin-T featuresGroup=Baseline Features, Protocol=Finetune2022.070.7990.808
TLVQMGroup=Existing Classical, Resolution Range=240P - 1080P2022.100.7990.803
Full-res Swin-TResolution Range=240P - 1080P, Features=Full-resolution Swin-T2022.100.7990.808
Proposed-LSTraining Setting=Mixed Databases, Scaling=Linear Re-scaling (LS)2021.080.78890.7806
CNN+VIDEVALGroup=Ensemble C+D, Protocol=Finetune2022.070.7850.81
CNN+VIDEVALGroup=Ensemble C+D, Resolution Range=240P - 1080P2022.100.7850.81
DisCoVQATrain on=KoNViD-1k2022.060.7820.797
FAST-VQA w/o VQ-representationsGroup=Ours, Protocol=Finetune2022.070.7820.803
VSFAGroup=Existing Fixed Deep, Protocol=Finetune2022.070.7730.795
VSFAGroup=Existing Fixed Deep, Resolution Range=240P - 1080P2022.100.7730.795
PVQ(w)Training Dataset=LSVQ, Evaluation Protocol=Cross-database2021.080.770.807
Patch-VQ (w/ patch)Training Set=LSVQ_train, GFLOPS=58501, CPU Time=539s, GPU Time=13.8s2022.110.770.807
RAPIQUEGroup=Existing Classical, Protocol=Finetune2022.070.7550.786
RAPIQUEGroup=Existing Classical, Resolution Range=240P - 1080P2022.100.7550.786
VIDEVALGroup=Existing Classical, Protocol=Finetune2022.070.7520.751
VIDEVALGroup=Existing Classical, Resolution Range=240P - 1080P2022.100.7520.751
PVQ(w/o)Training Dataset=LSVQ, Evaluation Protocol=Cross-database2021.080.7470.776
Patch-VQ (w/o patch)Training Set=LSVQ_train, GFLOPS=58501, CPU Time=539s, GPU Time=13.8s2022.110.7470.776
VSFATraining Dataset=LSVQ, Evaluation Protocol=Cross-database2021.080.7340.772
VSFATraining Set=LSVQ_train, GFLOPS=40919, CPU Time=466s, GPU Time=11.1s2022.110.7340.772
MDTVSFATraining Setting=Mixed Databases2021.080.72770.7784
KSVQETrain Dataset=KVQ2024.020.720.768
MDTVSFATrain on=KoNViD-1k2022.060.7160.759
CNN-TLVQMTrain on=KoNViD-1k2022.060.7130.752
ResNet-50Train set=KoNViD-1k2020.050.7—
GST-VQATrain on=KoNViD-1k2022.060.70.733
TLVQMTraining Dataset=LSVQ, Evaluation Protocol=Cross-database2021.080.670.691
TLVQMTraining Set=LSVQ_train, CPU Time=248s2022.110.670.691
DisCoVQATrain on=Youtube-UGC (236-fewer version)2022.060.6610.685
VIDEVALTraining Dataset=LSVQ, Evaluation Protocol=Cross-database2021.080.630.64
VIDEVALTraining Set=LSVQ_train, CPU Time=895s2022.110.630.64
VIDEVALTrain on=KoNViD-1k2022.060.6270.654
FastVQATrain Dataset=KVQ2024.020.6140.675
MDTVSFATrain on=Youtube-UGC2022.060.5820.603
VIDEVALTrain on=Youtube-UGC2022.060.5420.553
SimpleVQATrain Dataset=KVQ2024.020.5280.521
BRISQUETraining Dataset=LSVQ, Evaluation Protocol=Cross-database2021.080.5240.536
HOSATrain set=YouTube-UGC2020.050.49—