AI Paper Atlas

Research map / 3D vision / Face and shape

GPHM: Gaussian Parametric Head Model for Monocular Head Avatar Reconstruction

Yuelang Xu, Zhaoqi Su, Qiang Wu, Yebin Liu

TPAMI 2025 · 3D vision › Face and shape · approach: Gaussian and implicit head avatars · 15 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
initialize Gaussians from the guiding geometry model instead of FLAME (ablation)GPHM · Initialization ablationPSNR: 25.7 -> 28.0Improved
initialize Gaussians from the guiding geometry model instead of FLAME (ablation)GPHM · Initialization ablationLPIPS: 0.109 -> 0.085Improved
use 3D Gaussians instead of mesh (ablation)GPHM · Representation ablationPSNR: 15.7 -> 27.0Improved
add super-resolution network to mesh (ablation)GPHM · Representation ablationPSNR: 15.7 -> 17.3Improved
add super-resolution network to Gaussians (ablation)GPHM · Representation ablationPSNR: 27.0 -> 29.3Improved
fine-tune motion-related networks in addition to Gaussians (ablation)GPHMv2 · Self reenactmentPSNR: 28.1 -> 28.9Improved
add dynamic generator f_dyn (ablation)GPHMv2 · Self reenactmentLPIPS: 0.044 -> 0.041Improved
add dynamic generator f_dyn (ablation)GPHMv2 · Self reenactmentSSIM: 0.93 -> 0.94No clear effect
use GPHMv2 instead of FlashAvatarGPHMv2 · Self reenactmentPSNR: 28.8 -> 28.9Improved
use GPHMv2 instead of FlashAvatarGPHMv2 · Self reenactmentLPIPS: 0.051 -> 0.041Improved
use GPHMv2 fitting instead of PanoHeadGPHMv2 · Single image fittingPSNR: 29.1 -> 30.3Improved
add LivePortrait-synthesized expression condition images (ablation)GPHMv2 · Cross-identity reenactmentappearance leakage (qualitative)Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 5 of 5 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