AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Lip2Vec: Efficient and Robust Visual Speech Recognition via Latent-to-Latent Visual to Audio Representation Mapping

Yasser Abdelaziz Dahou Djilali, Sanath Narayan, Haithem Boussaid, Ebtessam Almazrouei, Merouane Debbah

ICCV 2023 · Vision-language and multimodal › Cross-modal learning · approach: Audio-visual source localization · 14 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace supervised CTC finetuning with latent-to-latent trainingAV-HuBERT Large (1759h pretrain, 433h) · LRS3 testWER: 38.6 -> 26.0Improved
replace supervised finetuning with latent-to-latent training on a Large encoder pretrained only on LRS3 433hAV-HuBERT Large (433h pretrain, 30h) · LRS3 testWER: 44.8 -> 55.4Hurt
use Lip2Vec instead of supervised CE finetuning on Base encoder (1759h, 433h)AV-HuBERT Base · LRS3 testWER: 34.8 -> 34.1Improved
add MSE loss term (alpha=0.01) to cosine loss (ablation)Lip2Vec · LRS3 test (Large, 1759h)WER: 34.6 -> 31.2Improved
increase MSE weight alpha beyond 0.05 (ablation)Lip2Vec · LRS3 test (Large, 1759h)WER: 31.2 -> 91.3 (alpha 0.5)Hurt
train longer without MSE loss (ablation)Lip2Vec alpha=0 · LRS3 test (Large, 1759h)WER: 31.4 vs 31.2No clear effect
use progressive audio-latent masking instead of no masking (ablation)Lip2Vec prior network · LRS3 test (Large, 1759h)WER: 75.2 -> 31.2Improved
use progressive masking instead of fixed 100% masking (ablation)Lip2Vec prior network · LRS3 test (Large, 1759h)WER: 65.3 -> 31.2Improved
add 30h VoxCeleb2-en to LRS3-30h for prior training (ablation)Lip2Vec Large · LRS3 testWER: 31.2 -> 30.4Improved
add 30h VoxCeleb2-en to LRS3-30h for prior training (ablation)Lip2Vec Large · VoxCeleb2-en testWER: 39.4 -> 33.1Improved
use CTC decoding with latent mapping instead of autoregressive decodersLip2Vec · VoxCeleb2-en inferenceruntime per 100 frames (s): 0.97 -> 0.07Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy