AI Paper Atlas

Research map / 3D vision / 3D generation

Bridging Facial Understanding and Animation via Language Models

Luchuan Song, Pinxin Liu, Haiyang Liu, Zhenchao Jin, Yolo Yunlong Tang, Zichong Xu, Susan Liang, Jing Bi, Jason J Corso, Chenliang Xu

CVPR 2026 · 3D vision › 3D generation · approach: Scene-aware human motion synthesis

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
train on Open3DFaceVid instead of MEAD (ablation)Motion2Language · Motion2LanguageCor_E (GPT-4): 3.03 -> 3.97Improved
train on Open3DFaceVid instead of YouTube (ablation)Motion2Language · Motion2LanguageCor_M (GPT-4): 3.92 -> 3.44Hurt
train on Open3DFaceVid instead of YouTube (ablation)Motion2Language · Motion2LanguageCor_E (GPT-4): 2.74 -> 3.97Improved
train on Open3DFaceVid instead of MEAD (ablation)Language2Motion · Language2MotionUSER: 2.88 -> 3.92Improved
train on Open3DFaceVid instead of MEAD (ablation)Language2Motion · Language2MotionFD: 32.14 -> 32.48Mixed
train on Open3DFaceVid instead of MEAD (ablation)Language2Motion · Language2MotionCor_E: 3.56 -> 3.90Improved
use geometry tokens in an LLM instead of image-token VLMsMotion2Language · Motion2Language (geometry input)Cor_E (GPT-4): 2.45 -> 4.02 vs Gemini2.5Improved
use geometry tokens vs Gemini on natural imagesMotion2Language · natural image inputCor_E (GPT-4): 4.21 (Gemini) vs 4.02 (Ours)Hurt
condition on LLM word-level prefix instead of GPT-style text conditioningLanguage2Motion · Language2MotionCor_E: 3.57 -> 4.13Improved
scale Motion2Language backbone beyond 4B-8B (ablation)Qwen3 backbones · Motion2Languagetoken accuracyNo clear effect
scale Language2Motion backbone from 0.7B to 3B (ablation)LLaMA backbones · Language2Motiontoken accuracyImproved

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