Viktor Larsson

@ducha_aiki My favorite result is table 10 by far. Excited about what we will be able to do with these uncertainties.

RoMa v2 results

RoMa v2: Harder Better Faster Denser Feature Matching @Parskatt et 11 al. tl;dr: in title. Predict covariance per-pixel, more datasets, use DINOv3, adjust architecture. arxiv.org/abs/2511.15706

Interested in doing research in 3D computer vision, geometry and machine learning? Open PhD position available in my group. Check out vlarsson.github.io/positions/ for more details on how to apply. Application deadline June 17th! Feel free to contact me if you have any questions.

Postdoc position available! Apply here lu.varbi.com/en/what:job/jo…

We have two new PhD student positions (Differentiable Neural Acoustics, Matrix factorization and optimization) at @lunduniversity - deadline 28 May 2024. Funded by WASP. Read more at kalleastrom.github.io. Apply at lu.varbi.com/en/what:job/jo…

🥳🥳Congrats to all the co-authors @visionviktor @Martin_R_Oswald @ZhaopengCui @AutoVisionGroup @mapo1 ! HUGE kudos to Zihan, who started working on this project during his bachelor's degree, and did such a wonderful job!

Interested in computer vision/ML/math/geometry? Open PhD positions available in my group, more details at vlarsson.github.io/positions/

@ducha_aiki 6/6 submitted. I feel like I spent both too much time and too little time on them...

CVPR reviews done. Just 4 this time

We have now released the full raw data of LaMAR, which include 3D lidar point clouds & meshes, on-device ToF depth maps, IMU data, GPS signals, etc. Huge potential to benchmark and train models for different geometric vision tasks! Check out lamar.ethz.ch to gain access

📢Excited to share "Privacy Preserving Localization via Coordinate Permutations", presented on #ICCV23. This is a joint work with Johannes Schönberger, @visionviktor, @mapo1🤗 For more info, check out 🌐Website: lpanaf.github.io/iccv23_privacy… 📄Paper: shorturl.at/jpzET

💣We are releasing the training code for GlueStick in a joint effort with our friends from LightGlue! With glue-factory you can train a matcher with any kind of feature, use different robust estimators, and evaluate in several benchmarks: ✅github.com/cvg/glue-facto…

If you are attending @ICCVConference and are interest in 3D vision, please join us at our tutorial on camera geometry estimation (sites.google.com/view/camgeo202…). The tutorial will start today (not tomorrow as in the printed program) at 13:30 in room W07.

📢I'm pleased to introduce GlueStick: A robust image matcher that combines points and lines📢 ➡️Website: iago-suarez.com/gluestick ➡️Code: github.com/cvg/GlueStick ➡️Paper: arxiv.org/pdf/2304.02008… This is a collaboration with Rémi Pautrat, Yifan Yu, @mapo1 and @visionviktor. 1/🧵

Excited to share that the popular Pixel-Perfect Structure-from-Motion #pixsfm is now on @condaforge. Installation is now as easy as "mamba install pixsfm" on Linux&MacOS. Note: #hloc and #colmap can be installed the same way. @pesarlin @ducha_aiki @prefix_dev #computervision 1/2

@B1ueber2y on the cover presenting our paper on 3D line mapping :)

@ducha_aiki @CVPR I agree. I also think oral presentations do better in workshops where the scope is a bit more narrow and there is more opportunity for discussion/questions.

I enjoyed “no-oral” mode of @CVPR with more keynotes and panel VERY much. Not sure what to do with 400 posters/halfday, but the oral program was amazing! #CVPR2023

Human-made environments contain a lot of straight lines, which are currently not exploited by most mapping approaches. With their recent work "3D Line Mapping Revisited" @B1ueber2y et al. take steps towards changing that. I visualized some of its key components with @rerundotio.

Interested in computer vision/ML/math/geometry? Fully-funded PhD positions available in my group, more details at vlarsson.github.io/positions/

3D Line Mapping Revisited @B1ueber2y, Yifan Yu, Rémi Pautrat, @mapo1 , @visionviktor tl;dr: Point-guided line SfM with scale-invariant scoring -> profit. Even improves SuperPoint + Colmap bundle adjustment arxiv.org/abs/2303.17504…

In case you missed our ECCV22 tutorial on localization and mapping for AR, you can find the short version here youtube.com/watch?v=5kH7FW… from a talk I gave last week at the Learning Machines seminar organized by @olofmogren at @RISESweden.

📢 Introducing LaMAR: a benchmark for Localization and Mapping with AR devices Website: lamar.ethz.ch Video: youtu.be/32XsRli2coo ➡️ Join our LaMAR tutorial at #ECCV2022 on Monday 24/10 - intro by Marc Pollefeys @mapo1 ⬇️ hours of AR data, 3 locations, over 2 years