jack_r_saunders Profile Banner
Jack Saunders Profile
Jack Saunders

@jack_r_saunders

Followers
600
Following
298
Media
116
Statuses
291

Talking about everything to do with Facial Avatars | PhD Student | Founder of @realsyncai

Bath, UK
Joined February 2020
Don't wanna be here? Send us removal request.
@jack_r_saunders
Jack Saunders
9 days
πŸ€– πŸ—žοΈ Free AI Generated Digital Humans Newsletter -> Like everyone else, I've been struggling to keep up with the sheer volume of papers and news in the Digital Human space. Over the past few weeks, I've developed an agentic (ish) AI pipeline to find and
Tweet media one
1
1
3
@jack_r_saunders
Jack Saunders
10 hours
Want to get up to three papers like this directly to your inbox and summarised by AI each day? You can sign up for free here:
0
0
0
@jack_r_saunders
Jack Saunders
10 hours
VisualSpeaker: Visually-Guided 3D Avatar Lip Synthesis. TLDR: Improves speech-driven facial mesh animation by rendering with 3DGS Avatars and running the results through a lip-reading network to get lip sync as a loss. πŸ“œ Paper:
Tweet media one
1
5
17
@jack_r_saunders
Jack Saunders
5 days
Want to get up to three papers like this directly to your inbox and summarised by AI each day? You can sign up for free here:
0
0
1
@jack_r_saunders
Jack Saunders
5 days
HyperGaussians: High-Dimensional Gaussian Splatting for High-Fidelity Animatable Face Avatars. TLDR: Lifts Gaussians to higher dimensions, which can then be conditioned for the problem of Gaussian Avatars. This enables much better reconstruction of high-frequency details and
1
27
100
@jack_r_saunders
Jack Saunders
6 days
πŸ€–πŸ“°Want to get up to three papers like this directly to your inbox and summarised by AI each day? You can sign up for free here:
0
0
0
@jack_r_saunders
Jack Saunders
6 days
ARIG: Autoregressive Interactive Head Generation for Real-time Conversations. TLDR: A method for dyadic conversations (speaking and listening) in real-time. This work uses an autoregressive diffusion MLP with added signals from context learning (by fusing LivePortrait parameters
2
43
203
@jack_r_saunders
Jack Saunders
8 days
Want to get up to three papers like this directly to your inbox and summarised by AI each day? You can sign up for free here:
0
0
1
@jack_r_saunders
Jack Saunders
8 days
SynMotion:Β Semantic-Visual Adaptation for Motion Customized Video Generation. TLDR: A DiT-based video generation model designed for motion retargeting. It uses a dual encoder architecture to separate subject and motion, and can be applied to non-human animals. πŸ“½οΈ Project Page:
1
6
45
@jack_r_saunders
Jack Saunders
12 days
3DGH: 3D Head Generation with Composable Hair and Face. TLDR: A Generative model of (static) Gaussian Avatars using a GAN. Here the head and hair are modelled separately with templates, and data is generated synthetically. πŸ“½οΈ Project Page: πŸ“œ Paper:
Tweet media one
1
20
72
@jack_r_saunders
Jack Saunders
14 days
πŸ’« Animate any rig using video diffusion models. Not a human-specific method, but really interesting as an idea. AnimaX: Animating the Inanimate in 3D with Joint Video-Pose Diffusion Models. TLDR: Animate any rig with an arbitrary skeleton using multi-view video diffusion
0
2
9
@jack_r_saunders
Jack Saunders
16 days
Controllable and Expressive One-Shot Video Head Swapping. TLDR: A reference-net-based approach for one-shot face swapping. This model works with landmark and background conditioning, with the former disentangled using a 3DMM. πŸ“½οΈ Project Page: πŸ“œ Paper:
3
28
117
@jack_r_saunders
Jack Saunders
19 days
One-shot Face Sketch Synthesis in the Wild via Generative Diffusion Prior and Instruction Tuning. TLDR: This work converts photorealstic portraits into sketches. Despite limited data, it is able to do this by optimising text instructions for a diffusion image-to-image model. πŸ“œ
Tweet media one
0
0
2
@jack_r_saunders
Jack Saunders
21 days
SyncTalk++: High-Fidelity and Efficient Synchronized Talking Heads Synthesis Using Gaussian Splatting. TLDR: A Gaussian-Splatting based person-specific talking head model which reduces the effect of OOD audio by training a discrete VQ-VAE over blendshapes. πŸ“½οΈ Project Page:
0
2
6
@jack_r_saunders
Jack Saunders
23 days
AlignHuman: Improving Motion and Fidelity via Timestep-Segment Preference Optimization for Audio-Driven Human Animation. TLDR: This DiT-based method first uses human preference on video for two categories, motion and fidelity. It then trains two LoRAs, one for each category,
0
8
47
@jack_r_saunders
Jack Saunders
24 days
πŸŽ‰ It's been an amazing weekend at #CVPR25, presenting our work GASP and meeting up with colleagues old and new! It's really crazy how many people and papers were there, but a few themes do seem to stand out for Head Avatars:. πŸ” Priors: A lot of works are starting to build
Tweet media one
0
1
16
@jack_r_saunders
Jack Saunders
24 days
RT @deblinaforAI: The visual computing group from @UniofBath πŸ‡¬πŸ‡§ at #CVPR25 in full force!.πŸ“ΈCo-chairing #CVPR25 publicity and .organising @W….
0
4
0
@jack_r_saunders
Jack Saunders
27 days
RT @SanyalSoubhik: It was fun to be a part of the organising team. In case you missed it, all the talks of the Digital humans symposium are….
0
1
0