Peter Wonka

Prof. of Computer Science at KAUST | Computer Vision, Machine Learning, Computer Graphics, Deep Learning

[Jiraphon Yenphraphai @JYenphraphai]

🧬 [1/3] Videos → dynamic 3D meshes 🧬

Excited to share Helix4D, our framework for complex 4D mesh generation from video!

It handles hard 4D cases: topology changes, inner surfaces, transparency, deformation, and thin structures.

👉 Project: https://t.co/xu0UZp5EID

[Gordon (Guocheng) Qian@CVPR2026 @guocheng_qian]

Excited to share that EasyV2V has been accepted to CVPR 2026! Data is the fuel. EasyV2V introduces a comprehensive and scalable video editing data pipeline. Curious how different types of data impact video editing performance? Check out our project page:
https://t.co/IsQkUoiVE2

[Ashkan Mirzaei @ashmrz10]

Meet our instructional video editing model, EasyV2V. Think nano-banana for videos. 🍌🎬

Turn simple instructions into powerful video edits.

Watch it in action 👉 https://t.co/emgEbdjVdo

[KAUST CEMSE @KAUST_CEMSE]

KAUST is hiring!

The Computer Science Program at #KAUST is inviting applications for faculty positions in Artificial Intelligence, with a focus on generative AI, #LLMs, or agentic AI, and in Large Scale High-Performance Computing.

With national investments driving massive growth in #AI infrastructure, data centers, and supercomputing, Saudi Arabia is creating one of the most advanced AI ecosystems in the world, and KAUST is driving this transformation through cutting-edge research, pioneering innovation, and cross-sector partnerships that connect discovery to real-world impact.

We welcome applications from across the spectrum of computer science, with current priorities in these fields.

Learn more and apply here:https://t.co/tDVvUQ6tRQ

#AgenticAI #HPC #FacultyPositions

[WindVChen @WindVChen]

📢V2M4 (Accepted by ICCV 2025)

We reconstruct 4D mesh animations from a single video with consistency in pose, appearance, topology, and texture.

project: https://t.co/nRbiSJXZSW
arxiv: https://t.co/VVAFUuZJwA

Collaboration with @_BiaoZhang , Xiangjun Tang and @peter_wonka

[Ashkan Mirzaei @ashmrz10]

[1/9] 🚀 We introduce 4Real-Video-V2, a method that can generate 4D scenes from a simple text prompt, viewable from any angle at any moment in time. It’s fast, photorealistic, and works on full scenes. Here's how it works and why it matters. 👇

https://t.co/vr9Wnvfx0s

[VCAI - MPI for Informatics @VcaiMpi]

Our paper, "Training-free Video Semantic Segmentation based on Diffusion Models," was accepted at CVPR 2025!

A great collaboration between KAUST and MPI.

Image diffusion models are powerful for image generation and serve as strong backbones with rich semantic understanding.

[Anton Obukhov @AntonObukhov1]

Update about the 4th Monocular Depth Estimation Workshop at #CVPR2025:
🎉 Website is LIVE! (link below)
🎉 Keynotes: Peter Wonka (@peter_wonka), Yiyi Liao (@yiyi_liao_), and Konrad Schindler
🎉 Challenge updates: new prediction types, baselines & metrics

[Mo @Mo_Li_KAUST]

Our preprint on deep learning based high-throughout human #blastoid classification is out It helps unbiased quality assessment of a large number of blastoids thx for Prof @peter_wonka and his team’s collaboration @KAUST_BESE @KaustResearch @biorxivpreprint https://t.co/4AUod27U67

[Matthias Niessner @MattNiessner]

📢📢 𝐏𝐫𝐄𝐝𝐢𝐭𝐨𝐫𝟑𝐃: 𝐅𝐚𝐬𝐭 𝐚𝐧𝐝 𝐏𝐫𝐞𝐜𝐢𝐬𝐞 𝟑𝐃 𝐒𝐡𝐚𝐩𝐞 𝐄𝐝𝐢𝐭𝐢𝐧𝐠 📢📢

We propose a training-free 3D shape editing approach that rapidly and precisely edits the regions intended by the user and keeps the rest as is.

Using a quickly brushed mask and a text prompt, we first apply multi-view editing in the 2D domain and then run our merging algorithm in the 3D feature space to ensure that the edited shape is loyal to the input shape.

Project Page: https://t.co/QRRcF1AP7Q
Video: https://t.co/pMOsbpYUKf

Great work by @ErkocZiya @cangumeli Chaoyang Wang @angelaqdai @peter_wonka @hyjameslee @PeiyeZ

[Zhenjun Zhao @zhenjun_zhao]

T-3DGS: Removing Transient Objects for 3D Scene Reconstruction

Vadim Pryadilshchikov, Alexander Markin, Artem Komarichev, @rusrakhimov, @peter_wonka, @burnaevevgeny

tl;dr: unsupervised transient detector and transient mask propagation

https://t.co/X5rjW2n9Df

[Biao Zhang @_BiaoZhang]

📢𝐆𝐞𝐨𝐦𝐞𝐭𝐫𝐲 𝐃𝐢𝐬𝐭𝐫𝐢𝐛𝐮𝐭𝐢𝐨𝐧𝐬
We propose to represent general geometries as distributions using diffusion models.

project site: https://t.co/4VDAoHGFM4
arxiv: https://t.co/VVFGiNT3zT
dataset: https://t.co/vrUPvy5EiS

Collaboration with Jing Ren and @peter_wonka

[Bing Li @summerie9]

Happy to share that our Vivid-Zoo paper has been accepted at #NeurIPS2024. The paper is collaborated with
@1zhengcheng
@WenxuanZhu75798
@_BiaoZhang
@peter_wonka
@BernardSGhanem

[Peter Wonka @peter_wonka]

@vincesitzmann @DavidCharatan @sizhe_lester_li @taiyasaki This was an insanely good talk at CVPR. Amazing. This is a great talk to learn from.

[KAUST @KAUST_News]

Join the #KAUST Global Fellowship Program, an incredible opportunity for exceptional postdocs with groundbreaking ideas to advance Saudi's research priorities. The program offers access to cutting-edge research tools, top-notch facilities and expert mentorship.

[Peter Wonka @peter_wonka]

LLM Blueprint: Enabling Text-to-Image Generation with Complex and Detailed Prompts
Hanan Gani (@hanan_shafi29), Shariq Farooq Bhat (@shariq_farooq), Muzammal Naseer, Salman Khan, Peter Wonka
ICLR 2024

https://t.co/9fSBvaO2wz

[Gordon (Guocheng) Qian@CVPR2026 @guocheng_qian]

Our image-to-3D generation work Magic123 gets accepted at #ICLR2024. It is a collaborated work among @KaustVision led by my supervisor @BernardSGhanem, my internship team at @Snapchat led by @SergeyTulyakov, @Eng_Hemdi from @Oxford_VGG, and @peter_wonka from @KAUST_News VCC.

[Peter Wonka @peter_wonka]

I collaborated with several colleagues from CAS and Sichuan on a survey on image and video inpaiting:
https://t.co/UXWof7k3sW

[Peter Wonka @peter_wonka]

Commencement at KAUST (⁦@cemseKAUST⁩ ⁦@isskoro⁩ ⁦@farnunea⁩ ⁦@_BiaoZhang⁩ and Wamiq and Yanze) Anna was selected as commencement speaker this year!

[Peter Wonka @peter_wonka]

A framework to advance SOTA in high-resolution depth estimation (@zhenyu_li9955 just joined as a new PhD student in August.)

[Zhenyu Li @zhenyu_li9955]

Check our new work PatchFusion! It estimates high-resolution metric depth with intricate details for high-resolution input image.

Zhenyu Li, Shariq Farooq Bhat (@shariq_farooq), Peter Wonka (@peter_wonka)

Webpage: https://t.co/KiRKNZvYX1