AI Paper Atlas

Research map / Video / Video generation

RoboGesture: Real-Time Semantic-aligned Co-Speech Gestures Generation for Humanoid Interaction

Zifan Wang, Ziang Ren, Pengyang Shi, Zirui Wang, Chenghuai Lin, Tianze Wang, Zekun Qi, Liangliang Zhao, HE WANG, Li Yi

ECCV 2026 · Video › Video generation · approach: Talking face generation · 1 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 semi-synthetic data (full 1000h vs none) (ablation)RoboGesture · Human study ablationSemantic Action Score: 4.281 -> 7.175Improved
use full semi-synthetic data instead of 1/4 (ablation)RoboGesture · Human study ablationSemantic Action Score: 6.316 -> 7.175Improved
add past motion context in cross-attention (ablation)RoboGesture · Human study ablationBeat Matching Score: 1.928 -> 7.529Improved
add FiLM injection of semantic tokens (ablation)RoboGesture · Human study ablationSemantic Action Score: 5.181 -> 7.175Improved
add semantic classification auxiliary task (ablation)RoboGesture · Human study ablationSemantic Action Score: 5.007 -> 7.175Improved
add Anti-Inertia CFG masking (15% past-motion drop) (ablation)RoboGesture · Human study ablationSemantic Action Score: 4.628 -> 7.175Improved
use flow-matching DiT instead of AR strategy (ablation)RoboGesture · Human study ablationSemantic Action Score: 4.843 -> 7.175Improved
add kinetic-aware loss (ablation)RoboGesture · Human study ablationHand Detail Score: 3.947 -> 7.203Improved
add MPC filter (ablation)RoboGesture · Human study ablationHuman-likeness: 6.891 -> 7.394Improved
add MPC safety filter (ablation)RoboGesture · Self-collision frame ratioself-collision (%): 4.16 -> 0.13Improved
add semantic-acoustic conditioning to the DiT (ablation)RoboGesture · Perturbation analysisHistory/Audio MSE ratio: 4.57 -> 2.04Improved
use RoboGesture instead of Semantic Gesticulator on diversityRoboGesture · BEATDIV: 0.2818 -> 0.2075Hurt

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