EmbedTalk: Talking Head Synthesis using Gaussian Embeddings

arXiv:2603.07604 · cs.CV · Submitted 2026-03-08 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: AI Radio. Generated commentary on the latest Artificial Intelligence papers.

Tom: I'm Tom.

Jane: And I'm Jane, here with Lu, Meng and our in-house large language model, Lalam.

Tom: We're looking at the paper: "EmbedTalk: Talking Head Synthesis using Gaussian Embeddings".

Jane: Here's the gist of the paper. Deformable 3D Gaussian Splatting (3DGS) has emerged as a popular method for real-time talking head synthesis, offering high-quality renderings at low latency.

Tom: EmbedTalk improves rendering quality, lip synchronisation, and motion consistency over previous 3DGS-based methods while remaining competitive with state-of-the-art generative models.

Jane: Replacing tri-plane features with embeddings also yields significantly more compact models that achieve 60+ FPS on a laptop GPU.

Tom: Synthesising audio-driven talking heads in real-time is an important task in computer vision, with applicability in film production, teleconferencing, and virtual assistants.

Title and authors: Tom: So we’re looking at this paper called EmbedTalk: Talking Head Synthesis using Gaussian Embeddings. It’s written by Arpita Saggar, Jonathan Darling, Duygu Sarikaya, and David Hogg. Jane, what’s the main takeaway from just looking at that title?

Jane: Well, it tells you right away that they are taking three dee Gaussian Splatting and adding these learnable embeddings to make talking heads <ref:2603.07604#pg1>. It suggests they are trying to fix some problems with how we currently do this kind of synthesis. It’s about making the three dee scene reconstruction more robust for speech input <ref:2603.07604#pg1>.

Tom: Exactly, it sounds like they're focusing on using these embeddings to model the facial movements driven by audio. I always like when papers move away from just using simple features and introduce something learnable that adapts to the data.

Lu: What’s interesting here is that they are specifically targeting the deformation process itself. They aren't just using a standard tri-plane encoding, which they say has issues with mirroring artifacts because of feature entanglement between subspaces.

Meng: So, if I understand this right, they are replacing those fixed tri-plane features with these per-Gaussian embeddings. That sounds like a way to make the model much more compact and flexible for handling different identities.

Lalam: I’m seeing here that they are aiming to generate stable talking heads with high audio visual alignment by using those embeddings for deformation, as you see on page two of the paper > EmbedTalk: Talking Head Synthesis using Gaussian Embeddings.

The paper's summary: Tom: So, what’s the core idea behind this approach? Basically, they are taking a dense reconstruction from COLMAP to start their three dee Gaussians and then giving each of those Gaussians its own set of learnable embeddings <ref:2603.07604#pg1>.

Jane: Right, so instead of every Gaussian being just a fixed shape, it gets an embedding that learns how it should deform based on the audio signal and the upper-face movements we feed in. It’s like giving each part of the three dee model a personality that reacts to what you’re saying <ref:2603.07604#pg1>.

Lu: What really stands out is their method for handling those high-frequency displacements, especially around the mouth region. They introduce positional encodings on top of these Gaussian embeddings to predict the deformations, which should help reconstruct those fine details more accurately.

Meng: From a deployment standpoint, having per-Gaussian embeddings means they are using smaller models for the deformation MLP instead of something huge that tries to model everything at once. That sounds good for real-time performance on consumer hardware.

Lalam: They’re feeding the speech signal and upper-face movements into this deformation MLP along with a positional encoding of the Gaussian embedding to predict things like changes in position and opacity, as mentioned on page three of EmbedTalk: Talking Head Synthesis using Gaussian Embeddings > Initialisation Per-Gaussian Embeddings Speech Signal Audio Encoder Δμ, Δα three deeGS Rasteriser + Dense Reconstruction e (blink, brow raise) Canonical three dee Gaussians (μ, s, r, α, f) a Camera Input Video Rendered Frames (Monocular) <ref:2603.07604#pg1,EmbedTalk: Talking Head Synthesis using Gaussian Embeddings>.

The paper's improvements: Tom: Okay, so how does this actually improve the existing methods? Page two mentions they are extending the embedding-based deformation paradigm to talking head synthesis. They claim this yields stable talking heads with high audio visual alignment.

Jane: And it goes further by using a local smoothness constraint on those embeddings to encourage similar embeddings for neighboring Gaussians, which should lead to better motion consistency overall. It addresses the wobbling issue that happens when you rely on imprecise facial tracking for camera pose, which they point out as a problem with prior work > EmbedTalk: Talking Head Synthesis using Gaussian Embeddings.

Lu: They also tackled the head wobbling by starting their Gaussians with a stable, dense reconstruction obtained via COLMAP, which is a solid initialization step. But they also used positional encodings specifically to reconstruct high-frequency displacements in the mouth region, which is key for lip synchronisation.

Meng: I’m interested in how much better the lip movement fidelity is compared to just using tri-plane features. If this works well, it means we can get much more realistic mouth movements without needing massive computational resources for every single feature interaction.

Lalam: They state that by generating mouth movements consistent with an identity’s style, their method achieves higher realism than state-of-the-art methods > EmbedTalk: Talking Head Synthesis using Gaussian Embeddings.

Conclusion: Tom: So, to wrap up, the main implication here is that learnable embeddings can significantly improve the quality of talking head synthesis while keeping the computational overhead down. They achieved better lip synchronisation and motion consistency compared to previous three deeGS-based works <ref:2603.07604#pg1>.

Jane: It seems like this approach lets us get much more faithful lip movements in these syntheses, especially when other methods often default to a closed-mouth state. It’s about getting the mouth right when you use audio input.

Lu: The method is restricted to facial animation for now, but it opens up possibilities for integrating this with full-body motion modelling techniques to support fully interactive avatars later on.

Meng: For deployment, they also started adaptive densification and pruning from iteration five hundred onwards, which shows they have a structured way of managing the memory footprint as the model gets bigger.

Lalam: And just to be clear about what it doesn't do, the paper says this method is focused on facial animation and isn't necessarily aiming for full-body motion modeling yet > EmbedTalk: Talking Head Synthesis using Gaussian Embeddings.

School of Computer Science, University of Leeds · Leeds Institute of Medical Education, School of Medicine, University of Leeds

cs.CV

Submitted: 2026-03-08

Updated: 2026-11-03

Comments: Camera Ready version for WACV 2027

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 77/100

The gist: Deformable 3D Gaussian Splatting (3DGS) has emerged as a popular method for real-time talking head synthesis, offering high-quality renderings at low latency.

Key concepts

Deformable 3D Gaussian Splatting (3DGS)
A technique used to represent 3D scenes using a collection of small, volumetric shapes called Gaussians. These Gaussians are rendered by projecting them onto a 2D plane. In this method, the shape of these Gaussians can be dynamically changed or 'deformed' based on input data like speech and motion.
Per-Gaussian Embeddings (zg)
These are small, learnable numerical codes assigned to each individual Gaussian in the 3D scene. Instead of using fixed features, these embeddings allow the system to capture unique characteristics for every part of the head. These embeddings are crucial because they enable the model to understand and accurately deform specific facial attributes like position and opacity based on audio input.
Deformation Mechanism
This is how EmbedTalk changes a Gaussian's shape. It works by projecting the Gaussian primitives into continuous, lower-dimensional subspaces. These subspaces are then combined with audio signals and control signals, which are finally projected back into the 3D space to create an attribute-wise deformation that accurately reflects the desired facial movement.

Terminology

Summary

Deformable 3D Gaussian Splatting (3DGS) has emerged as a popular method for real-time talking head synthesis, offering high-quality renderings at low latency. p. 0

Where the paper says it

20 MB 61 FPS Figure 1. EmbedTalk models speech deformations in talking head synthesis with learnable per-Gaussian embeddings, resulting in more accurate mouth movements and reduced computational overhead compared to tri-plane-based methods Abstract Deformable 3D Gaussian Splatting (3DGS) has emerged as a popular method for real-time talking head synthesis, offering high-quality renderings at low latency.

EmbedTalk improves rendering quality, lip synchronisation, and motion consistency over previous 3DGS-based methods while remaining competitive with state-of-the-art generative models. p. 0
Where the paper says it

We introduce EmbedTalk, which leverages these embeddings to model speech-driven facial deformations for talking head synthesis. Comprehensive experiments show that EmbedTalk improves rendering quality, lip synchronisation, and motion consistency over previous 3DGS-based methods, while remaining competitive with state-of-the-art generative models.

Replacing tri-plane features with embeddings also yields significantly more compact models that achieve 60+ FPS on a laptop GPU. p. 0
Where the paper says it

Comprehensive experiments show that EmbedTalk improves rendering quality, lip synchronisation, and motion consistency over previous 3DGS-based methods, while remaining competitive with state-of-the-art generative models. Replacing tri-plane features with embeddings also yields significantly more compact models that achieve 60+ FPS on a laptop GPU (RTX 2060 6 GB).

Introduction to the Problem

Synthesising audio-driven talking heads in real-time is an important task in computer vision, with applicability in film production, teleconferencing, and virtual assistants. Methods dominated by imagebased generative models typically do not provide pose control and are slow to render. p. 0

Where the paper says it

In recent years, methods for synthesising talking heads have been dominated by imagebased generative models [2, 20, 29, 45, 46] and radiance fields [12, 15, 23, 25, 33]. Whilst generative models produce high resolution results, they typically do not provide pose control, are slow to render, and can generate uncharacteristic movements which reduce realism and applicability.

3D Gaussian Splatting (3DGS) has emerged as the preferred choice due to its fast rendering speed and low memory requirements compared to neural radiance fields (NeRFs). p. 0
Where the paper says it

Whilst generative models produce high resolution results, they typically do not provide pose control, are slow to render, and can generate uncharacteristic movements which reduce realism and applicability. Within radiance fields, 3D Gaussian Splatting (3DGS) [19] has emerged as the preferred choice due to its fast rendering speed and low memory requirements, compared to neural radiance fields (NeRFs) [30].

Tri-plane encodings suffer from mirroring artefacts caused by feature entanglement between subspaces. p. 0
Where the paper says it

[10, 12, 16, 24, 25, 34]. However, tri-plane representations suffer from mirroring artefacts caused by feature entanglement between subspaces [22].

Methodology Overview

EmbedTalk leverages per-Gaussian embeddings to model speech-driven facial deformations for talking head synthesis. The process begins with a dense reconstruction obtained from COLMAP to initialise the 3D Gaussians. p. 3

Where the paper says it

SfM-based initialisation provides better reconstruction than random initialisation [19]. Consequently, we initialise the Gaussians for EmbedTalk using a dense reconstruction obtained from COLMAP, downsampled to ≤ 100K points (to prevent out of memory errors).

Each Gaussian is associated with a learnable embedding zg, and for each frame, the corresponding speech signal a and upper-face movements e are fed into the deformation MLP. p. 2
Where the paper says it

Each Gaussian is also associated with a learnable embedding zg. For each frame, the corresponding speech signal a and upper-face movements e are fed into the deformation MLP, along with a positional encoding of zg to predict the Gaussian deformations (∆µ, ∆α).

The final deformed Gaussians are passed to the rasteriser along with the viewing direction to render the head onto the combined torso and scene background. p. 2
Where the paper says it

For each frame, the corresponding speech signal a and upper-face movements e are fed into the deformation MLP, along with a positional encoding of zg to predict the Gaussian deformations (∆µ, ∆α). The deformed Gaussians are passed to the rasteriser, along with the viewing direction (camera), to render the head onto the combined torso and scene background.

Deformation Mechanism Details

The deformation is achieved by projecting one or more Gaussian primitives to a set of continuous, lower dimensional subspaces. This lower dimensional representation is then combined with the audio embedding and control signals and projected back into the Gaussian space to produce an attributewise deformation with respect to the canonical representation. p. 3

Where the paper says it

jected back into the Gaussian space to produce an attributewise deformation with respect to the canonical representation. The reprojection is achieved using a decoder multilayer perceptron (MLP) as shown below: Gcanonical = {µ, r, s, f,α} (5) {∆µn, ∆rn, ∆sn, ∆fn, ∆αn} = MLP(H(Gcanonical) ⊕ an ⊕ en) (6) G n deform = {µ + ∆µn, r + ∆rn, s + ∆sn, f + ∆fn, α + ∆αn} (7) where H is a multi-resolution tri-plane encoder [10, 24, 25, 33] and ⊕ denotes concatenation.

Specifically, the deformation module is a shallow MLP with a separate prediction head for each deformed attribute (µ and α). p. 3
Where the paper says it

We provide this information to the deformation MLP through a positional encoding of the feature set E, comprising six facial action units [13]. Our embedding-based deformation is given by: Gcanonical = {µ, r, s, f,α, z} (8) {∆µn, ∆αn} = MLP(γ(z) ⊕ an ⊕ γ(en)) (9) G n deform = {µ + ∆µn, r, s, f, α + ∆αn} (10) The deformation module for EmbedTalk is a shallow MLP with a separate prediction head for each deformed attribute (µ and α).

The method chooses to deform only position and opacity because facial animation primarily involves changes in motion and visibility. p. 3
Where the paper says it

Unlike contemporary methods that deform all Gaussian attributes [3, 10], or attributes related to size and orientation [12, 24, 25], we choose to deform only position and opacity. This selection is grounded in the fact that facial animation primarily involves changes in motion (head tilting, mouth opening and closing) and visibility (teeth/tongue appearing or disappearing).

Training Strategy and Optimization

The training process involves joint optimisation of the canonical Gaussian attributes, the per-Gaussian embeddings, and the deformation module to minimise the rendering loss L1. A local smoothness constraint is applied to encourage similar embeddings for nearby Gaussians given by Lemb reg. p. 3

Where the paper says it

It therefore follows that nearby Gaussians should also have similar embeddings to promote motion consistency. Motivated by prior work [3, 28], we apply a local smoothness constraint to encourage similar embeddings for nearby Gaussians, given by: Lemb reg = 1 k|G| X i∈G X j∈KNNi;k (wi,j∥zgi − zgj∥2) (11)

The total loss for the setup is given by Ltotal = L1(vn, vˆn) + λfaceLLP IP S(vn, vˆn) + λmouthLLP IP S(v m n, vˆ m n) + Lemb reg + λopaLopacity. p. 4
Where the paper says it

Finally, we minimise the mean of the Gaussian opacities to mitigate floaters. The total loss for our setup is given by: Ltotal = L1(vn, vˆn) + λfaceLLP IP S(vn, vˆn) +λmouthLLP IP S(v m n , vˆ m n ) + Lemb reg +λopaLopacity (12) where vn ∈ V represents the ground truth frame, vˆn is the predicted frame, v m n denotes the ground truth frame cropped to the mouth region, and vˆ m n is the prediction cropped to the mouth region.

Adaptive control of Gaussians through densification and pruning begins from iteration 500 and continues till iteration 49000. p. 3
Where the paper says it

During training, the Gaussian count is controlled through adaptive densification and pruning strategies [19]. Rendering: By default, 3DGS renders Gaussians on a single colour background image, which is not ideal for our approach.

Experimental Results and Comparisons

EmbedTalk performs best on all rendering metrics and achieves the most temporally consistent motion in the self-driven setting. EmbedTalk also performs best for identity-specific lip synchronisation (LMD) and has the highest Sync-C score among 3DGS-based methods enabled by positional encodings. p. 4

Where the paper says it

EmbedTalk (COLMAP init) performs best on all rendering metrics and achieves the most temporally consistent motion. Our method also performs best for identity-specific lip synchronisation (LMD) and has the highest Sync-C score among 3DGS-based methods, enabled by positional encodings that capture high-frequency details.

The computational costs show that EmbedTalk yields compact models that achieve over 60 FPS on a laptop GPU. p. 5
Where the paper says it

Computational costs for the 3DGS-based methods. EmbedTalk yields compact models that achieve over 60 FPS on a laptop (RTX 2060 6GB) GPU.

Qualitatively, EmbedTalk generates the most faithful lip movements while other methods often default to a closed-mouth state. p. 5
Where the paper says it

5 shows comparisons with generative methods. Among Gaussian Splatting-based approaches, EmbedTalk generates the most faithful lip movements while other methods often default to a closed-mouth state.

Conclusion and Future Work

EmbedTalk shows how learnt embeddings can be leveraged to deform 3D Gaussians for talking head synthesis producing high quality talking heads with accurate and realistic lip synchronisation. The method is restricted to facial animation and could be integrated with full-body motion modelling techniques to support fully interactive avatars. p. 7

Where the paper says it

While we anticipate that EmbedTalk would generalise to new videos with diverse emotions and expressions, this remains to be experimentally validated. Additionally, our method is restricted to facial animation and could be integrated with full-body motion modelling techniques to support fully interactive avatars.

The paper advocates for the use of explicit labelling and watermarking techniques to help distinguish real videos from AI-generated videos. p. 7
Where the paper says it

However, our method carries potential for misuse through generation of deepfakes for identity theft or misinformation. We advocate for the use of explicit labelling and watermarking techniques to help distinguish real videos from AI-generated videos.

The source code will be placed in the public domain on acceptance. p. 0
Where the paper says it

Replacing tri-plane features with embeddings also yields significantly more compact models that achieve 60+ FPS on a laptop GPU (RTX 2060 6 GB). Our code will be placed in the public domain on acceptance.

References

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. p. 8

Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

p. 8
Where the paper says it

References [1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting.

In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026. p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

p. 8
Where the paper says it

Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

[1] Madhav Agarwal, Mingtian Zhang, Laura Sevilla-Lara, and Steven McDonagh. Gaussianheadtalk: Wobble-free 3d talking heads with audio driven gaussian splatting. In IEEE/CVF Winter Conference on Applications of Computer Vision 2026, 2026.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved Talking Head Synthesis for High Fidelity Motion: EmbedTalk can produce the most faithful lip movements compared to other methods, which often default to a closed-mouth state. This allows the system to generate talking heads with accurate and realistic lip synchronisation and captures high-frequency details through positional encodings.

  2. Reduced Computational Overhead for Real-Time Application: By replacing tri-plane features with embeddings, EmbedTalk achieves significantly more compact models that achieve 60+ FPS on a laptop GPU. This makes the system viable for real-time use on consumer hardware, as it is nearly twice as fast as other methods in inference speed.

  3. Enhanced Motion Consistency and Stability: The method incorporates a local smoothness constraint to encourage similar embeddings for neighbouring Gaussians, which promotes motion consistency. Furthermore, the use of positional encodings enables embeddings to disentangle motion discontinuous from smooth deformations, leading to less shakiness in renderings.

  4. Superior Visual Realism in Cross-Driven Scenarios: While generative models produce exaggerated movements, EmbedTalk scores highly on user studies for video realism and image quality. This suggests that the system can generate talking heads that are less like it had been AI-generated compared to state-of-the-art generative baselines.

  5. Efficient Resource Management in Model Training: The training strategy employs adaptive densification and pruning strategies starting from iteration 500, which allows for the control of Gaussian count while maintaining reconstruction stability. This provides a structured approach to managing the memory footprint and computational cost during the lengthy training process.

Abstract

Deformable 3D Gaussian Splatting (3DGS) has emerged as a popular method for real-time talking head synthesis, offering high-quality renderings at low latency. Tri-planes are a common choice for encoding Gaussians prior to deformation since they provide a compact and continuous representation. However, tri-plane encodings are limited by grid resolution and approximation errors introduced by projecting 3D volumetric fields onto 2D subspaces. Recent work has demonstrated the effectiveness of per-Gaussian embeddings for driving temporal deformations in 4D scene reconstruction. We introduce EmbedTalk, which leverages these embeddings to model speech-driven facial deformations for talking head synthesis. Comprehensive experiments show that EmbedTalk improves rendering quality, lip synchronisation, and motion consistency over previous 3DGS-based methods, while remaining competitive with state-of-the-art generative models. Replacing tri-plane features with embeddings also yields significantly more compact models that achieve 60+ FPS on a laptop GPU (RTX 2060 6 GB). Code and further details are available at https://arpita2512.github.io/EmbedTalk/

Related papers