Lipreading On Lrs3 Ted

Word Error Rate (WER)

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

比較表

モデル名	Word Error Rate (WER)
syncvsr-data-efficient-visual-speech	31.2
syncvsr-data-efficient-visual-speech	21.5
sub-word-level-lip-reading-with-visual	30.7
auto-avsr-audio-visual-speech-recognition	19.1
where-visual-speech-meets-language-vsp-llm	25.4
visual-speech-recognition-for-multiple	31.5
learning-audio-visual-speech-representation-1	26.9
end-to-end-audio-visual-speech-recognition	43.3
audio-visual-representation-learning-via	26.2
large-scale-visual-speech-recognition	55.1
discriminative-multi-modality-speech	57.8
conformers-are-all-you-need-for-visual-speech	12.8
spatio-temporal-fusion-based-convolutional	60.1
asr-is-all-you-need-cross-modal-distillation	59.8
es3-evolving-self-supervised-learning-of	37.1
jointly-learning-visual-and-auditory-speech	23.4
relaxed-attention-for-transformer-models	25.51
unified-speech-recognition-a-single-model-for	21.5
recurrent-neural-network-transducer-for-audio	33.6
es3-evolving-self-supervised-learning-of	40.3
deep-audio-visual-speech-recognition	58.9
unified-speech-recognition-a-single-model-for	22.3
sub-word-level-lip-reading-with-visual	40.6