3D Computer Vision at Apple. I like 3D vision and training neural networks.
Joined July 2012
- Tweets2.7K
- Following251
- Followers1.4K
- Likes3.3K
Flexibility, along with robustness, is what make models widely applicable in practice.
Looking forward to an age of SfM naver failing.
To match or not to match?
FlexGlue 🐙, our new image-matching model, can match every pixel AND reject doppelganger pairs, tackling a common failure case in SfM.
➡️State-of-the-art SfM and matching, blazing fast
➡️Code: github.com/cvg/FlexGlue
➡️Paper: arxiv.org/pdf/2610.32976
1/
Just speechless.
Berlin-based search engine Ecosia, used by government services, is dropping its French AI partner, Mistral, its CEO told POLITICO.
The company will now bet on open-source models, including those from China.
vist.ly/5mh5y
1
7
1,287
If your paper contains a bunch of Claudeisms I'm going to assume that you don't know what you're doing.
Goated 🇸🇪🤝🇫🇷
Today we release Poincar3, our newest 3D model trained entirely through SSL on RGB image sequences.
I find the coolest thing is that, without labels, it learns multi-view geometry. Just using the attention map, reliable tracks can be formed (see below).
davnords.com/poincar3
It's a must-s33
We improved the MUSt3R demo again, it runs on my laptop at ~20fps and has loop-closure (with memory updates!). If you'd like to try this, come see our #ECCV2026 demo Friday morning, with online reconstruction+panoptic segmentation, and other cool works from our lab!
6
1,400
