AI Paper Atlas

Research map / 3D vision / Face and shape

CodeTalker: Speech-Driven 3D Facial Animation With Discrete Motion Prior

Jinbo Xing, Menghan Xia, Yuechen Zhang, Xiaodong Cun, Jue Wang, Tien-Tsin Wong

CVPR 2023 · 3D vision › Face and shape · approach: Speech-driven 3D facial animation · 183 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
use a speaker-agnostic motion codebook rather than a shape-entangled codebook (ablation)CodeTalker · VOCA-TestReconstruction Error: 2.75 -> 0.08Improved
use a speaker-agnostic motion codebook rather than a shape-entangled codebook (ablation)CodeTalker · BIWI-Test-AReconstruction Error: 4.07 -> 2.83Improved
use a speaker-agnostic motion codebook rather than a shape-entangled codebook (ablation)CodeTalker · BIWI-Test-ALip Vertex Error: 6.41 -> 4.79Improved
add Instance Normalization during self-reconstruction learning (ablation)CodeTalker · VOCA-TestReconstruction Error: 0.12 -> 0.08Improved
add Instance Normalization during self-reconstruction learning (ablation)CodeTalker · BIWI-Test-AReconstruction Error: 3.27 -> 2.83Improved
replace CodeTalker's code-query framework with token prediction supervised only by categorical cross-entropy (ablation)CodeTalker · BIWI-Test-ALip Vertex Error: 4.7914 -> 9.6356Hurt
replace CodeTalker's code-query framework with token prediction using cross-entropy and feature regularity losses (ablation)CodeTalker · BIWI-Test-ALip Vertex Error: 4.7914 -> 5.1138Hurt
replace CodeTalker's code-query framework with token prediction using cross-entropy, feature regularity, and motion losses (ablation)CodeTalker · BIWI-Test-ALip Vertex Error: 4.7914 -> 5.0254Hurt

Later papers that compared with this method

7 later accepted papers report CodeTalker as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.

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