List your company

Audio-visual speech separation on LRS2 2Mix

15.9SDRiTDFNet-large
4.3567.35310.3513.347Jan 25, 2024

Evaluation results

MethodLinks
TDFNet-largeParams (M)=6.5, MACs (G)=47.22024.0115.915.83.210.949
TDFNet (MHSA + Shared)Params (M)=4.2, MACs (G)=38.62024.0115.2153.160.938
CTCNetParams (M)=7.1, MACs (G)=167.22024.0114.614.33.080.931
TDFNet-smallParams (M)=5.8, MACs (G)=152024.0113.713.63.10.931
CaffNet-C*Notes=Variant of CaffNet-C2024.0112.5—1.150.89
VisualVoiceParams (M)=77.82024.0111.811.53—
Thanh-Dat2024.0111.6—3.1—
Conv-TasNetParams (M)=5.62024.0110.710.3——
A-FRCNNParams (M)=6.32024.0110.19.4——
CaffNet-C2024.0110—0.940.88
SuDORM-RFParams (M)=2.72024.019.59.1——
uPITParams (M)=92.72024.014.83.6——