AI Paper Atlas

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

Efficient Audio-Visual Speech Separation with Discrete Lip Semantics and Multi-Scale Global-Local Attention

Kai Li, Gao Kejun, Xiaolin Hu

ICLR 2026 · Vision-language and multimodal › Cross-modal learning · approach: Audio-visual source localization · 3 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
add global attention (GA) to the GLA block (ablation)Dolphin separator · LRS2SI-SNRi (dB): 15.6 -> 16.8Improved
add local attention (LA) to the GLA block (ablation)Dolphin separator · LRS2SI-SNRi (dB): 15.9 -> 16.8Improved
add both GA and LA to a plain depthwise-conv block (ablation)Dolphin separator · LRS2SI-SNRi (dB): 10.4 -> 16.8Improved
use HDA layer instead of 1D large-kernel convolution (ablation)Dolphin · LRS2SI-SNRi (dB): 16.5 -> 16.9Improved
use discrete (VQ) lip features instead of continuous autoencodervideo encoder · LRS2SI-SNRi (dB): 15.2 -> 16.8Improved
add audio-aligned semantic distillation to LipCoder (DP-LipCoder)video encoder · LRS2SI-SNRi (dB): 16.3 -> 16.8Improved
replace 3D ResNet-18 with DP-LipCodervideo encoder · LRS2SI-SNRi (dB): 17.0 -> 16.8No clear effect
replace original video encoder with DP-LipCoder in IIANetIIANet · LRS2SI-SNRi (dB): 16.0 -> 15.4Hurt
replace original video encoder with DP-LipCoder in CTCNetCTCNet · LRS2SI-SNRi (dB): 14.3 -> 14.0Hurt
use early audio-visual fusion in the separator encoder instead of later fusion (ablation)Dolphin · LRS2SI-SNRi (dB): +0.6 dBImproved
add VQ module to the video encoder (ablation)DP-LipCoder · LRS2SI-SNRi (dB): approximately 0.5 dBImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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