Excited to share our #CVPR2021 work LoFTR. LoFTR can extract high-quality semi-dense matches even in indistinctive regions with low-textures, motion blur, or repetitive patterns. The project page is at: zju3dv.github.io/loftr/.

Reconstructing 3D Human Pose by Watching Humans in the Mirror pdf: arxiv.org/pdf/2104.00340… abs: arxiv.org/abs/2104.00340 project page: zju3dv.github.io/Mirrored-Human/

I have tried LoFTR for my datasets and it is pure magic. The longer post is going soon, just wanted to say "Wow!" Img-pair1: thermal vs visible + viewpoint change. Img-pair2: light+scale P.S. I just realised that LofTR can stand also for "Lord OF the Rings". Is it intentional?

As one of my closest mentors and #rolemodel, extremely honoured to be in Kostas Daniilidis's top 5 cited papers (with top 4 in close reach). Thank you Kostas' for your continued support! Georgios Pavlakos Xiaowei Zhou

Amazing LoFTR is in #kornia master branch now! 🥳 Implement state-of-the are image matching methods with 2 lines of code Jiaming Sun ❤️ #CVPR2021 #opensource #computervision #DeepLearning

#3DGV Our next speaker is Xiaowei Zhou (Zhejiang). Talk title: "Towards More Accessible 3D Digitalization of Humans and Scenes". Link: youtube.com/watch?v=_G4RLB… Host: Yang Liu (MSRA) Panelists: Yebin Liu (THU) and Lan Xu(ShanghaiTech) 10/27 7:00 pm (PDT)

Thank you Qixing Qixing Huang@ICCV 2025 for organizing this fantastic series of 3DGV seminars.

We revisit the old Manhattan-world constraint in the new era of neural scene representations. Check out our #CVPR2022 oral paper on Neural 3D Scene Reconstruction with the Manhattan-world Assumption: zju3dv.github.io/manhattan_sdf/

No CAD models, no instance or category-specific training, all you need is a strong generic feature matcher like SuperGlue or LoFTR. Check out our CVPR 2022 paper on object pose estimation without CAD models!

Building Brandenburg Gate in hours from Internet photos

Exciting work from Yiming Xie

Check out our SIGGRAPH paper on modeling multiple closely interacting people with few cameras!

Excited to share our new work on real-time 4K rendering of volumetric video

Excited to introduce EasyVolcap, an easy-to-use library to help you develop your own volumetric video systems!

🚀 Excited to share our breakthrough paper "SpatialTracker: 3D Space Tracking for 2D Pixels" - selected as highlight paper at #CVPR2024! We lifted the dense pixels tracking into 3D space👇 For more details, welcom to check out: henry123-boy.github.io/SpaTracker/

Efficient LoFTR code released! Try it out to get semi-dense matching in real-time: github.com/zju3dv/Efficie…

Tired of short multi-view video datasets? Check out our new SelfCap dataset with up to 10 minutes of 24-cameras high-quality dense view recording at 4K resolution! Data released for our Long Volumetric Video paper: zju3dv.github.io/longvolcap Dataset link: forms.gle/MzJqZjBfyZ53fR…

Excited to share our work MatchAnything: We pre-train strong universal image matching models that exhibit remarkable generalizability on unseen multi-modality matching and registration tasks. Project page: zju3dv.github.io/MatchAnything/ Huggingface Demo: huggingface.co/spaces/LittleF…

🚀 We release SpatialTrackerV2: the first feedforward model for dynamic 3D reconstruction and 3D point tracking — all at once! Reconstruct dynamic scenes and predict pixel-wise 3D motion in seconds. 🔗 Webpage: spatialtracker.github.io 🔍 Online Demo: huggingface.co/spaces/Yuxihen…

🚀 Join our Volumetric Video Reconstruction Challenge! 🏆 Top prize of $2500 for the winner.