AI Paper Atlas

Research map / 3D vision / Face and shape

FlexAvatar: Flexible Large Reconstruction Model for Animatable Gaussian Head Avatars with Detailed Deformation

Cheng Peng, Zhuo Su, Liao Wang, Chen Guo, Zhaohu Li, Chengjiang Long, Zheng Lv, Jingxiang Sun, Chenyangguang Zhang, Yebin Liu

CVPR 2026 · 3D vision › Face and shape · approach: Gaussian and implicit head avatars · 3 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 UV position map as driving signal to the UNet (ablation)FlexAvatar · FaceCap ablationPSNR: 22.51 -> 23.32Improved
add UV position map as driving signal to the UNet (ablation)FlexAvatar · FaceCap ablationLPIPS: 0.1845 -> 0.1797Improved
add data distribution adjustment (ablation)FlexAvatar · FaceCap ablationPSNR: 22.74 -> 23.32Improved
add data distribution adjustment (ablation)FlexAvatar · FaceCap ablationLPIPS: 0.1868 -> 0.1797Improved
add mouth-region LPIPS loss (ablation)FlexAvatar · FaceCap ablationPSNR: 23.10 -> 23.32Improved
add mouth-region LPIPS loss (ablation)FlexAvatar · FaceCap ablationSSIM: 0.8890 -> 0.8895Improved
add 10-second test-time refinement to feed-forward modelFlexAvatar · NeRSemble self reenactmentPSNR: 21.1516 -> 22.6313Improved
add 10-second test-time refinementFlexAvatar · NeRSemble self reenactmentLPIPS: 0.2193 -> 0.1833Improved
use FlexAvatar feed-forward instead of best baseline GAGAvatarFlexAvatar · NeRSemble self reenactmentAKD: 3.8730 -> 3.6518Improved
use FlexAvatar feed-forward instead of Portrait4D-v2FlexAvatar · NeRSemble self reenactmentPSNR: 19.6279 -> 21.1516Improved
use refined FlexAvatar instead of HeadGAPFlexAvatar · self reenactmentPSNR: 20.77 -> 22.6313Improved
add training identities (ablation)FlexAvatar feed-forward · FaceCapPSNRImproved

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