Audio-Visual Speech Recognition

Benchmarks

Dataset Name	SOTA Method	Metric
LRS3 clean (test)	MMS-LLAMA	WER0.72	70	4d ago
LRS2 (test)	USR 2.0	WER1.3	34	4d ago
LRS-3 Babble noise at 0dB SNR (test)	LP Conformer	WER1.9	32	4d ago
LRS3 (test)	Auto-AVSR	WER0.9	18	2d ago
WildVSR (test)	USR 2.0	WER0.385	12	4d ago
LRS2 (clean)	MIR-GAN	WER2.2	12	4d ago
LRS3 (test)	GER w/ Auto-AVSR	Overall Score43	10	2d ago
LRS2 50% visual occlusion (test)		WER (Overall)6.4	10	2d ago
TED LRS3	VGG CONFORMER	WER0.009	10	4d ago
LRS3	Llama-AVSR	WER0.008	9	4d ago
MuAViC Noise environment (test)	XLAVS-R 2B	Accuracy (En)49.5	9	3d ago
MuAViC Clean environment (test)	XLS-R 300M	En Acc2.5	9	3d ago
LRS3 noisy	AV-HuBERT + CMA + MoHAVE	Average Error Rate4.2	8	2d ago
LRS3 Pixelated face	CAV2vec	WER (Babble, -10 dB)26	7	2d ago
LRS3 Occlusion by hands	CAV2vec	WER (Babble, -10 dB)26.6	7	2d ago
LRS3 Object occlusion and noise	CAV2vec	WER (Babble, -10 dB)25.8	7	2d ago
LRS3 noisy synthesized using MUSAN noise (test)	MIR-GAN	WER5.6	7	4d ago
MuAViC (test)	AV-HUBERT	Accuracy (Ara)89.4	7	4d ago
LRS2 noisy (MUSAN)	MIR-GAN	WER7	6	4d ago
LRS3 + DEMAND Object Occlusion + Noise (test)	CAV2vec	Error Rate (PARK)2.8	5	2d ago
FLEURS Noise environment (test)	XLAVS-R 2B	WER74	5	3d ago
FLEURS Clean environment (test)	XLAVS-R 300M	WER32.5	5	3d ago
LRS-3 Average Babble of noisy SNR levels (-10 to 5dB) (test)	UADF	WER9.9	5	4d ago
LRS-3 Babble noise at -5dB SNR (test)	UADF	WER10.7	5	4d ago
LRS-3 Babble noise at -10dB SNR (test)	UADF	WER21.8	5	4d ago

Showing 25 of 32 rows