AI Paper Atlas

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

C3ASD: Multi-Level Consistency-Driven Representation Learning for Robust Active Speaker Detection

Jin Hong, Jisoo Park, Junseok Kwon

ECCV 2026 · Vision-language and multimodal › Cross-modal learning · approach: Self-supervised audio-visual representation learning · 0 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 inter-modality consistency loss (ablation)C3ASD (Light-ASD) · AVA valmAP: 93.61 -> 93.70Improved
add intra-modality track-aware contrastive loss (ablation)C3ASD (Light-ASD) · AVA valmAP: 93.61 -> 93.62No clear effect
add prediction-level consistency loss (ablation)C3ASD (Light-ASD) · AVA valmAP: 93.61 -> 93.68Improved
combine all three consistency losses (ablation)C3ASD (Light-ASD) · AVA valmAP: 93.61 -> 93.80Improved
add consistency losses under object occlusionC3ASD (Light-ASD) · AVA val corruptedmAP: 76.86 -> 78.90Improved
add consistency losses under pixelationC3ASD (Light-ASD) · AVA val corruptedmAP: 93.15 -> 93.47No clear effect
add consistency losses under MUSAN/DEMAND audio noiseC3ASD (Light-ASD) · AVA val corruptedmAP: 90.2 -> 90.7 (DEMAND)Improved
add consistency losses under joint occlusion and MUSAN noiseC3ASD (Light-ASD) · AVA val corruptedmAP: +1.23% average mAPImproved
add inter-modality loss: audio-visual cosine similarity (ablation)C3ASD (Light-ASD) · AVA valcosine similarity: 0.555 -> 0.834Improved
add prediction loss under SNR -10 dB noise (ablation)C3ASD (Light-ASD) · AVA valAV/V prediction disagreement: 0.324 -> 0.162Improved
add consistency losses on cross-dataset transferC3ASD (Light-ASD) · WASDmAP: 85.3 -> 86.1Improved
add intra-modality loss without cross-modal anchoring (appendix PCA) (ablation)C3ASD · AVA val embeddingsmean paired distance: 0.276 -> 0.886Hurt

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