Zirui Wang

Postdoc and DPhil @AVLOxford

[Zirui Wang @ziruiwang_]

Excited to share that our Oxford Day-and-Night dataset has been accepted to the NeurIPS 2025 Dataset track! See you in San Diego!

📦 Dataset: https://t.co/FGOKTaKv5W
🌐 Project Page: https://t.co/VDbGYgKq5z

[Zhenjun Zhao @zhenjun_zhao]

Seeing in the Dark: Benchmarking Egocentric 3D Vision with the Oxford Day-and-Night Dataset

@ziruiwang_, @wenjing_bian, @li_xinghui, @Yifu_Tao, Jianeng Wang, @MauriceFallon, Victor Adrian Prisacariu

tl;dr: egocentric dataset with GT 3D and lighting variation for benchmarking NVS and visual relocalization

https://t.co/jq2KGfVQlz

[Zirui Wang @ziruiwang_]

Finally someone proves 3D is still useful in robotic models!

[Yu Xiang @YuXiang_IRVL]

Just in time: the Best Paper Award in Robot Learning #ICRA2026 uses 3D camera pose to improve policy learning.

Pretty straightforward: robots live in a 3D world.

Image credit: @CSProfKGD

[Zirui Wang @ziruiwang_]

@sharecashBTC @mondorobotics same question here!

[Zirui Wang @ziruiwang_]

What a refreshing read!

[Zhi (Leo) Wang @TX_Leo_Wang]

1/ 🧠Humans are the best robot data source!
2/ 👓Human egocentric video is rich in quantity, but poor in quality.
3/ Beyond scaling data, smarter representation and architecture matter just as much.
4/ Want an open-source framework to train your own learn-from-human-data robot policy?

🚀We introduce HumanEgo: Zero-Shot Robot Learning
from Minutes of Human Egocentric Videos⬇️

✦ Zero-Shot Human-to-Robot Transfer
✦ Robot-Data-Free
✦ Just 30 min of data per task
✦ Collect by Anyone, Anytime, Anywhere
✦ Deploy on Any Robot, Any Camera, Any Environment
✦ Open-Source & Easy-to-Implement

Let's squeeze every bit of signal out of human data!

🌐 Website: https://t.co/JfsW8x6wtq
📄 Paper: https://t.co/tsaIiatmNi
💻 Code: https://t.co/jZjghCcjh2
📹 Video: https://t.co/QWmJmQ9GgQ
🧵 1/n

[Jing Wu 🍄 @jingwu23]

Excited to share that our synthetic evaluation benchmark is now integrated into @Voxel51.

Huge thanks to Harpreet Sahota and the @Voxel51 team for making this happen!

Check it out 👉: https://t.co/NY4zzckZTt

[Jing Wu 🍄 @jingwu23]

📣 2 New Papers!!

I am happy to share my 2 recent works. 🥳

1️⃣ [ArXiv Preprint]: ⏱️OS-Marathon: Benchmarking Computer-Use Agents on Long-Horizon Repetitive Tasks
- 🌐Project: https://t.co/NORAd1ptuL
- 📑 Paper: https://t.co/UxERGlevQb

2️⃣ [3DV] Reflect3r: Single-View 3D Stereo Reconstruction Aided by Mirror Reflections
- 🌐Project: https://t.co/9X66b0dIrH
- 🚀Code: https://t.co/Jv3JxE2aER
- 🤗HF: https://t.co/VEji2srAru

[Jing Wu 🍄 @jingwu23]

📣📣 Introducing Reflect3r @3DVconf
✅ Single-view to 3D reconstruction by using mirror reflections as an auxiliary view
✅ A synthetic evaluation dataset containing 16 fully customizable Blender scenes.
This is a joint work with @ziruiwang_, Iro Laina, and @viprad

🌐 Project Page: https://t.co/9X66b0dIrH
📄 Paper: https://t.co/nk5NWyKOUv
👩‍💻 Code: https://t.co/Jv3JxE2aER
🤗 Huggingface: https://t.co/VEji2srAru

[RuiningLi @RayLi234]

Introducing Particulate: a feed-forward model for 3D object articulation 💻✂️👓🧳

Particulate gives you a fully articulated 3D object, including part segmentation, kinematic structure & motion constraints, in a single forward pass in ~10secs.

🏅SOTA performance!
💡GenAI compatible: Turns AI-generated 3D meshes into fully articulated models!

Project page: https://t.co/8yYFpYdEkY
Code: https://t.co/CUuubxqbdY

[Zirui Wang @ziruiwang_]

@gabriberton Thanks, glad you enjoyed it! We were also impressed by how well MegaLoc handled those extreme lighting conditions the first time we saw it!

[Zirui Wang @ziruiwang_]

@AkashKumbar6573 @zhenjun_zhao @ysbhalgat @RayLi234 @viprad Ah thanks for the note, should be fixed now!

[Zhenjun Zhao @zhenjun_zhao]

Active View Selector: Fast and Accurate Active View Selection with Cross Reference Image Quality Assessment

@ziruiwang_, @ysbhalgat, @RayLi234, @viprad

tl;dr: CrossScore helps active NVS

https://t.co/HXSZHVYtWd

[Zhenjun Zhao @zhenjun_zhao]

Seeing in the Dark: Benchmarking Egocentric 3D Vision with the Oxford Day-and-Night Dataset

@ziruiwang_, @wenjing_bian, @li_xinghui, @Yifu_Tao, Jianeng Wang, @MauriceFallon, Victor Adrian Prisacariu

tl;dr: egocentric dataset with GT 3D and lighting variation for benchmarking NVS and visual relocalization

https://t.co/jq2KGfVQlz

[Zirui Wang @ziruiwang_]

Glad to see a couple of our methods featured (and getting beaten🤕) in this video 😅

[Two Minute Papers @twominutepapers]

NVIDIA Unveils AI For 150x Faster 3D Modeling!
Full video: https://t.co/rEglRiYFSM

[Zirui Wang @ziruiwang_]

Check out our diffusion based 3D object detection. Will be an oral presentation in 3DV 2025!

https://t.co/Y0bnSkv7Dk

#3dv #3dv2025

[Wenjing Bian @wenjing_bian]

CatFree3D: Category-agnostic 3D Object Detection with Diffusion

TL;DR: A diffusion-based 3DOD pipeline boosting accuracy & generalization, plus a new evaluation metric

📄 https://t.co/GAkEhUGYPT

w/ @ziruiwang_, Andrea Vedaldi

Look forward to the oral presentation at #3DV2025

[Zirui Wang @ziruiwang_]

Just found a great paper Fast3R, processing 1000 images in Dust3R style in one go!

https://t.co/91Is4oP4so

Btw, training takes a week on 64 A100 😅 and inferencing 1500 images takes 5min on 1 A100

[Zirui Wang @ziruiwang_]

I'm excited that our paper got into ICLR 2025!

Great work done by @DavidLA05031686!

This is another great work on camera localisation from our group @AVLOxford (other works are mostly from @ShuaiC8 and @wenjing_bian)

[Yash Bhalgat @ysbhalgat]

📢 Paper accepted to #ICLR2025 🎉

"GSLoc: Efficient Camera Pose Refinement via 3D Gaussian Splatting"

TL;DR: a novel test-time camera pose refinement framework leveraging 3DGS as the scene representation and MASt3R for 2D matching.

🔗: https://t.co/zK99FUQhCy

[Jiao Sun @sunjiao123sun_]

Mitigating racial bias from LLMs is a lot easier than removing it from humans!

Can’t believe this happened at the best AI conference @NeurIPSConf

We have ethical reviews for authors, but missed it for invited speakers? 😡

[Eleonora Svanberg @EleonoraSberg]

I'm a PhD student at @UniofOxford and I think I'm living in a fairytale :-)

Foxes playing around in the snow at Magdalen College this morning — absolutely magical!

[Zirui Wang @ziruiwang_]

@Parskatt My hypothesis is that using CrossScore as a loss function requires it to be a smooth manifold, and with <400 scenes trained so far, I don’t believe the manifold is smooth enough for GD. We plan to work on it in future.

[Zirui Wang @ziruiwang_]

Come to our poster #297 this afternoon! #ECCV

[Zirui Wang @ziruiwang_]

Thanks for sharing :)

TLDR: Introducing an image evaluation method that assesses rendered images via images from other viewpoints, eliminating the need for ground truth.

Primary use: assessing novel views in NVS tasks.

⛳️ https://t.co/y0QdBq1ngl
📄 https://t.co/yDBmW0fuqR

[Zirui Wang @ziruiwang_]

@Parskatt We focus on measuring instead of back propagation or GD in this work.
We did try with GD, the results weren’t great.