AI Paper Atlas

Research map / Video / Video generation

Contrastive Decoupled Representation Learning and Regularization for Speech-Preserving Facial Expression Manipulation

Tianshui Chen, Jianman Lin, Zhijing Yang, Chunmei Qing, Yukai Shi, Liang Lin

IJCV 2025 · Video › Video generation · approach: Talking face generation · 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
integrate CDRL supervision into NEDNED · MEAD intra-IDaverage FAD: 2.108 -> 1.026Improved
integrate CDRL supervision into NEDNED · MEAD intra-IDaverage CSIM: 0.831 -> 0.914Improved
integrate CDRL supervision into NEDNED · MEAD cross-IDaverage LSE-D: 9.906 -> 9.351Improved
integrate CDRL into NED without retraining CDRLNED · RAVDESS cross-IDaverage FAD: 5.412 -> 4.816Improved
add CCRL content supervision (ablation)Ours (NED) · MEAD cross-IDLSE-D: 9.883 -> 9.351Improved
add audio as content prior in CCRL (ablation)CCRL · MEAD cross-IDLSE-D: 9.812 -> 9.351Improved
add emotion-aware negatives in CCRL contrastive loss (ablation)CCRL · MEAD cross-IDLSE-D: 9.382 -> 9.351Improved
add CERL emotion supervision (ablation)Ours (NED) · MEAD intra-IDCSIM: 0.837 -> 0.914Improved
add CERL emotion supervision (ablation)Ours (NED) · MEAD cross-IDCSIM: 0.774 -> 0.792Improved
use dot product instead of cross-attention in CERL (ablation)CERL · MEAD intra-IDCSIM: 0.899 -> 0.914Improved
filter emotionally expressive images with CLIP for CERL training (emotion-augmented contrastive learning) (ablation)CERL · MEAD intra-IDCSIM: 0.888 -> 0.914Improved
combine ASCCL with CDRL (ablation)CDRL on NED · MEAD cross-IDFAD: 4.344 -> 4.252Improved

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