One Basis to Animate Them All: Gaussian Blendshape Distillation for Real-Time Avatars

Ramazan Fazylov; Stamatis Lefkimmiatis; Ivan Laptev

cs.CVOct 1, 2026

One Basis to Animate Them All: Gaussian Blendshape Distillation for Real-Time Avatars

Authors: Ramazan Fazylov, Stamatis Lefkimmiatis, Ivan Laptev

Organizations: Mohamed bin Zayed University of Artificial Intelligence · MWS AI

Abstract

3D Gaussian avatars support fast rendering, however, their real-time animation is often challenged by the costly neural inference. We address this bottleneck and show that the animation of pretrained avatar models can be closely approximated by a linear combination of identity-independent blendshapes. Building on this finding, we introduce GALA (Gaussian Animation via Linear Approximation), a distillation method that replaces per-frame heavy neural decoding with a shallow coefficient predictor and a linear blend. To improve fidelity and reduce memory requirements, we propose to construct the basis using block-local PCA under a rendering-aware metric and a memory budget. Our method learns a shallow MLP network to predict blendshape coefficients and applies to various animation architectures without retraining original models. We validate GALA by accelerating the inference of three distinct avatar models for 3D animation of facial expressions and full-bodies with clothing dynamics. Across these models, our distillation generalizes to held-out identities and reduces CPU animation cost by up to three orders of magnitude while preserving most of the rendering quality. Excellent results of our method confirm the shared linear structure of learned avatar representations and enable highly efficient and accurate animation at frame rates reaching up to 60fps on mobile devices. Project page: https://ramazan793.github.io/gala/

Figures & tables

Figure 1: GALA enables efficient, high-fidelity animation across diverse Gaussian avatar models. By introducing shared blendshapes and a lightweight coefficient predictor, GALA accelerates CPU animation by up to three orders of magnitude and enables real-time deployment on mobile devices.

	Photo- real.	Identity- shared	Linear at run time	Where the basis comes from
3D morphable models ( Li et al., 2017b )	$\times$	✓	✓	hand-built from scans
Per-person blendshapes ( Zielonka et al., 2025 )	✓	$\times$	✓	fitted per subject
Fixed-rig avatars ( He et al., 2025 )	✓	✓	✓	chosen up front
Animatable avatar models ( Yu et al., 2025 )	✓	–	$\times$	none (a network)
GALA (Ours)	✓	✓	✓	discovered in a trained model

Table 1: Comparison of avatar animation representations. Our method extracts an identity-shared blendshape basis from a pretrained nonlinear avatar model, complementing approaches based on predefined rigs or per-subject fitting. The extracted basis closely approximates the model’s learned deformations and supports efficient linear blending for unseen identities.

Figure 2: Overview of GALA. Residuals

r_{i}

of the host model

A

with respect to the neutral avatar

A(w_{i},\theta_{0})

(Section 3.2.1 ) are decomposed by a spatially local, rendering-aware PCA under a memory budget

\mathcal{B}

into the block-diagonal basis

U

(Section 3.2.2 ). A coefficient network

f_{\phi}

is trained to regress the projected coefficients

c^{\star}

(Section 3.2.3 ). At inference, the neutral avatar is computed once per subject, and every frame costs one evaluation of a shallow network

f_{\phi}

and a linear blend.

	FID $\downarrow$	AED $\downarrow$	AED-jaw $\downarrow$	ID $\uparrow$	APD $\downarrow$	FPS $\uparrow$	mFPS $\uparrow$
AGORA	3.17	0.682	0.021	0.75	0.025	250 ∗	1
EG3D	3.28	–	–	–	–	–	–
GGHEAD	4.06	–	–	–	–	–	–
Next3D	3.18	0.930	0.046	0.74	0.031	15 ∗	–
GAIA	3.85	0.530	0.040	0.72	0.027	52 ∗	–
AGORA + GALA (ours)	3.47	0.705	0.022	0.76	0.026	1,275	60

Table 2: Comparison to the state of the art on the benchmark of each host model. In every panel the first row is the host model and the last row is the same model distilled by our method (+GALA). The rows in between are the methods compared in the paper of each host model, with the numbers reported there. The AGORA row is taken from Fazylov et al. (2025) . The FlexAvatar and DynaAvatar rows are measured by us on the authors’ released checkpoints, following their evaluation protocols (Appendix F ). Best and second-best results, excluding the host model. EG3D and GGHead are not animatable. FPS is measured on a desktop GPU with a precomputed neutral avatar and mFPS on a mobile phone in a web browser. ∗ As reported in the model’s paper.

Figure 3: Qualitative comparison with the host models ( Original ), two examples per host model. Ours (projected) uses projected coefficients $c^{\star}$ , which require running the host model and show what the basis can express (Section 3.2.3 ), and Ours the coefficients predicted by the coefficient network. Error maps show the absolute difference to the host model per pixel, averaged over the RGB channels (values in $[0,1]$ ), from 0 (black) to 0.2 (yellow). Zoom in for details.

Per-attribute	Rendering-aware		PSNR to host $\uparrow$
bases	PCA	allocation	AGORA	Flex	Dyna
$\times$	$\times$	$\times$	32.77	21.56	21.38
$\times$	✓	✓	37.83	29.48	25.86
✓	$\times$	✓	42.32	32.74	34.04
✓	✓	$\times$	37.23	30.39	21.34
✓	✓	✓	42.57	33.09	34.62

Table 3: Ablation of the basis construction. Row 1 of (a) is simple PCA (Section 3.2.1 ).

Rank	AGORA	Flex	Dyna
Global, $K_{a,j}=K$	36.39	31.91	26.17
Per group, $K_{a,j}=K_{a}$	39.19	32.97	33.95
Per group and block, $K_{a,j}$ (ours)	42.57	33.09	34.62

Table 4: Allocation of components, PSNR to the host model.

Figure 4: Comparison to alternative distillation: quality against the CPU time of the animation step, lower is better on both axes. Teacher is the host model, our models are labeled with the components per Gaussian and the basis size, and the diamond marks our final model.

Appendix figures & tables13 assets

Supplementary material from the paper’s appendix.

Appendix

Panel of Figure 5 — Figure 5: Our browser demos on a mobile phone, screenshots from the video on our project page . Left: AGORA + GALA driven by the tracked expressions of a video. Right: five FlexAvatar + GALA avatars, each reconstructed from one photo, around a table (top), and five DynaAvatar + GALA full-body avatars with clothing dynamics (bottom). The overlays show the frame rate and the timing of the animation step.

	AGORA	FlexAvatar	DynaAvatar
Gaussians $n$	228,693	58,361	40,000
attribute groups $a$	position, scale, rotation	position, scale, rotation, opacity, color	position, scale, rotation, opacity, color, skinning weights
animation parameters $\theta$	FLAME expression and jaw	FLAME-derived expression code	current and 15 preceding SMPL-X poses
spatial blocks $B$	32	32	32
metric $\alpha$ , $\beta$ , $\gamma$	1, 16, 0	1, 16, 24	1, 64, 0
bytes per value $b$	2	2	4

Appendix

Table 5: Host models and the settings of our distilled models. Components are listed in total and per attribute group, in the order of the attribute groups. The animation step is timed on a CPU without rendering. For DynaAvatar, both the host model and ours additionally pose the Gaussians by linear blend skinning, which takes 10 ms per frame.

	AGORA	FlexAvatar	DynaAvatar
identity descriptor $w$	latent of the neutral render (512)	embedding of the fitted avatar code (256)	identity tokens and shape parameters of the host encoder (3,082)
animation parameters $\theta$	FLAME expression, jaw and eyelids (55)	articulation code (135)	16 SMPL-X frames $\times$ 79 values
hidden layers	256, 512	512, 512	per frame 512, 512, two temporal convolutions of 512, then 1,024, 1,024
activation	leaky ReLU	GELU	SiLU
outputs $K$	1,016	4,860	2,686
parameters	0.80 M	2.96 M	15.2 M

Appendix

Table 6: Coefficient networks of the distilled models of Table 2 . Hidden layers are listed with their widths. Parameters and multiply-accumulate operations (MAC) are those of one animation step of the network, without the basis. The FlexAvatar identity encoder (6.4 M parameters) runs once per identity and is not included.

	AGORA (FFHQ)					FlexAvatar (VFHQ)					DynaAvatar (4D-Dress)
	PSNR ${}^{\ast}\uparrow$	SSIM ${}^{\ast}\uparrow$	FID $\downarrow$	ID $\uparrow$	AED $\downarrow$	PSNR $\uparrow$	SSIM $\uparrow$	LPIPS $\downarrow$	CSIM ${}^{\dagger}\uparrow$	AED ${}^{\dagger}\downarrow$	PSNR $\uparrow$	SSIM $\uparrow$	LPIPS $\downarrow$	PSNR ${}^{\ast}\uparrow$
Simple PCA	28.82	0.894	3.50	0.797	0.792	21.41	0.799	0.119	0.75	0.271	22.59	0.956	0.069	21.27
GALA (ours)	31.18	0.929	3.47	0.761	0.705	22.31	0.818	0.109	0.68	0.245	23.22	0.957	0.066	23.87

Appendix

Table 7: Simple PCA and our full method with predicted coefficients on the benchmarks of Table 2 , at the memory budgets of Table 5 . FlexAvatar is evaluated in self-reenactment on VFHQ, with the enrollment for both models. Best results. ∗ With respect to the host model. † Cross-reenactment. The higher ID and CSIM of simple PCA reflect under-animation, since a static head obtains a CSIM of 0.813 on VFHQ.

	student	parameters	size	CPU	FID / LPIPS $\downarrow$
AGORA , deformation branch with 512 / 256 / 128 channels at $64^{2}$ / $128^{2}$ / $256^{2}$
A1	64 / 32 / 16 channels	0.56 M	2.2 MB	9.4 ms	4.14
A2	192 / 96 channels, no $256^{2}$ block	1.03 M	4.1 MB	12.2 ms	3.73
A3	128 / 64 / 32 channels	0.89 M	3.6 MB	16.2 ms	3.71
A4	256 / 128 / 64 channels	1.77 M	7.1 MB	45.1 ms	3.71
A5	448 / 224 / 112 channels, pruned from the host	3.60 M	14.4 MB	124.4 ms	3.60

Appendix

Table 8: Distilled networks of Figure 4 . For AGORA, channels are listed per resolution of the deformation branch.

L

and

d

are the number and the width of the transformer blocks, and

N

the number of attended Gaussians for DynaAvatar. Size is the storage of the student in single precision. CPU times are for the animation step at batch size one with eight threads on an Intel Core Ultra 9 285K, without the skinning of DynaAvatar. Quality is the vertical axis of Figure 4 : FID on FFHQ for AGORA and LPIPS with respect to the ground truth for FlexAvatar and DynaAvatar.

	Input per subject			Self reenactment					Cross reenactment
				training frames		held-out frames
	Views	Frames	Fit	PSNR $\uparrow$	LPIPS $\downarrow$	PSNR $\uparrow$	LPIPS $\downarrow$	AED $\downarrow$	CSIM $\uparrow$	AED $\downarrow$	APD $\downarrow$
INSTA
RGBAvatar	1	2,790	55 s	33.21	0.0487	30.23	0.0518	0.173	0.524	0.898	0.0499
FlexAvatar †	1	1	$\sim$ 1 min	24.58	0.0858	24.43	0.0675	0.311	0.612	0.700	0.0219
GALA (ours)	1	1	$\sim$ 1 min	23.76	0.1152	23.80	0.0812	0.387	0.639	0.800	0.0234

Appendix

Table 9: Comparison to per-subject blendshape avatars on 6 held-out identities of INSTA and 3 identities of Ava256. Input per subject: views, frames per view and time of the fit or of our enrollment, without face tracking. CSIM uses the input photo as reference. Best held-out and cross results, excluding the host model ( † ).

	Input per subject			Self reenactment, held-out frames			Cross reenactment
	Views	Frames	Fit	PSNR $\uparrow$	LPIPS $\downarrow$	AED $\downarrow$	CSIM $\uparrow$	AED $\downarrow$	APD $\downarrow$
RGBAvatar	1	127	65 s	20.70	0.1084	0.251	0.451	0.860	0.0645
FlexAvatar †	1	1	$\sim$ 1 min	19.72	0.1363	0.292	0.628	0.634	0.0188
GALA (ours)	1	1	$\sim$ 1 min	19.15	0.1496	0.325	0.666	0.787	0.0212

Appendix

Table 10: Comparison to RGBAvatar on 3 short clips of the VFHQ test set. RGBAvatar is fitted to the first 70% of each clip, FlexAvatar and our model use the first frame, and self reenactment is scored on the last 20%. Best results, excluding the host model ( † ).

Figure 6: Comparison to RGBAvatar on INSTA and Ava256. RGBAvatar is fitted to the video (INSTA) or to the 15-view capture (Ava256) of the subject, whereas the host model FlexAvatar and GALA use one photo. The self rows show held-out frames next to the ground truth, the cross rows are driven by VFHQ videos of other people (left) without ground truth.

Figure 7: Comparison to RGBAvatar on a VFHQ clip of Table 10 . RGBAvatar is fitted to the first 70% of the clip, whereas the host model FlexAvatar and GALA use its first frame. Rows: a training frame of RGBAvatar, a held-out frame, and cross reenactment driven by a VFHQ video of another person (left).

Figure 8: Comparison to GEM on an Ava256 identity. GEM is fitted to the 15-view capture of the subject, whereas the host model FlexAvatar and GALA use one photo. Rows: a training frame of GEM from one of its training cameras, a held-out frame from the camera of the input photo, and cross reenactment driven by a VFHQ video of another person (left).

Panel of Figure 9 — Figure 9: More difficult examples, with the error scale of Figure 3 . Neutral only is the neutral avatar with all coefficients zero, Projected uses projected coefficients $c^{\star}$ and GALA the predicted ones. Rows 1 and 2: held-out AGORA identities at the frame with the largest residual of the host model. Rows 3 and 4: held-out Ava256 subjects in the few-shot and the single-shot setting. Row 5: the frame with the largest error of our model in the DynaAvatar sequence of Figure 3 . In row 4 the coefficient network fails: 23.6 dB PSNR with respect to the host model with predicted coefficients, 40.2 dB with projected ones.

Panel of Figure 10 — Figure 10: Local motions learned by our blendshapes. The first column highlights the region affected by the visualized blendshapes, and the other columns show two states of their motion for three identities: closing of an eye (AGORA), closing of the mouth (FlexAvatar) and lifting of the front part of an open garment (DynaAvatar).

Panel of Figure 11 — Figure 11: Further local motions, in the layout of Figure 10 : a smile (AGORA), lowering of a brow together with the upper eyelid (FlexAvatar, cropped to the brow) and the opposite front part of the garment (DynaAvatar).

Explore similar work

AGORA: Adversarial Generation Of Real-time Animatable 3D Gaussian Head Avatars
Dec 6, 2025Ramazan Fazylov, Sergey Zagoruyko, Aleksandr Parkin +2Animatable 3D Human Avatars Avatars
Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, UAE · Polynome AI, Dubai, UAE · MTS AI, Moscow, Russia
High-Fidelity Mobile Avatars with Pruned Local Blendshapes
May 3, 2026Youyi Zhan, He Wang, Tianjia Shao +1Animatable 3D Human Avatars High-Fidelity Avatar
State Key Lab of CAD&CG, Zhejiang University · University College London
Self-Learning Expression Deformations for Data-Efficient Gaussian Avatars
Jun 4, 2026Jiahao Yang, Xiaohang Yang, Qing Wang +3Animatable 3D Human Avatars Facial Expression Recognition
Queen Mary University of London

Dec 6, 2025, cs.CVJ/K move · Enter open · S save

AGORA: Adversarial Generation Of Real-time Animatable 3D Gaussian Head Avatars

Ramazan Fazylov, Sergey Zagoruyko, Aleksandr Parkin, Stamatis Lefkimmiatis+1

Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, UAE · Polynome AI, Dubai, UAE · MTS AI, Moscow, Russia

Animatable 3D Human Avatars Avatars Feed-Forward 3D Gaussian Reconstruction 3D Object Understanding Generative Adversarial Network

Open paper AskPDF arXiv

The generation of high-fidelity, animatable 3D human avatars remains a core challenge in computer graphics and vision, with applications in VR, telepresence, and entertainment. Existing approaches based on implicit representations like NeRFs suffer from slow rendering and dynamic inconsistencies, while 3D Gaussian Splatting (3DGS) methods are typically limited to static head generation, lacking dynamic control. We bridge this gap by introducing AGORA, a novel framework that extends 3DGS within a generative adversarial network to produce animatable avatars. Our key contribution is a lightweight, FLAME-conditioned deformation branch that predicts per-Gaussian residuals, enabling identity-preserving, fine-grained expression control while allowing real-time inference. Identity is further preserved through spatial shape conditioning of the identity branch, and expression fidelity is enforced via a dual-discriminator training scheme leveraging synthetic renderings of the parametric mesh. AGORA generates avatars that are not only visually realistic but also precisely controllable. Quantitatively, we outperform state-of-the-art NeRF-based methods on expression accuracy while rendering at 250 FPS on a single GPU and, notably, at

\sim

9 FPS under CPU-only inference -- to our knowledge the first demonstration of CPU-only animatable 3DGS avatar synthesis. This work represents a significant step toward practical, high-performance digital humans. Project website: https://ramazan793.github.io/AGORA/

Abstract

Explore similar work

AGORA: Adversarial Generation Of Real-time Animatable 3D Gaussian Head Avatars

High-Fidelity Mobile Avatars with Pruned Local Blendshapes

Self-Learning Expression Deformations for Data-Efficient Gaussian Avatars