Tomasz Malisiewicz

Sunset at #WACV2024, day three Shoutout to all the new #computervision researchers I met over the last three days, and thanks for coming to our Meta Reality Labs booth to learn about projectaria.com

#WACV2024 Sunset on the Big Island. Come say hi tomorrow at the Project Aria booth to learn how we use Aria glasses for egocentric research at Reality Labs @ Meta.

@CSProfKGD I’ll be there! See you on the Big Island

Looking forward to attending #WACV2024 next week! 😎

Quest 3 arriving soon! I’m going to play some recroom paintball this weekend!

@vincesitzmann Great work! Interesting avenue of thought for those still working on bottom-up 3D reconstruction and trying to apply machine learning directly to 3D representations.

Introducing “Diffusion with Forward Models”, 𝗮 𝗺𝗼𝗱𝗲𝗹 𝘁𝗵𝗮𝘁 𝗰𝗮𝗻 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗲 𝗱𝗶𝘃𝗲𝗿𝘀𝗲, 𝗿𝗲𝗮𝗹 𝟯𝗗 𝘀𝗰𝗲𝗻𝗲𝘀 𝗳𝗿𝗼𝗺 𝗮 𝘀𝗶𝗻𝗴𝗹𝗲 𝗶𝗺𝗮𝗴𝗲, 𝘁𝗿𝗮𝗶𝗻𝗲𝗱 𝘄𝗶𝘁𝗵 𝗶𝗺𝗮𝗴𝗲𝘀 𝘄/𝗼 𝗮𝗻𝘆 𝟯𝗗 𝗱𝗮𝘁𝗮! …ffusion-with-forward-models.github.io 1/n

#CVPR2023 in Vancouver was a lot of fun! My first conference since becoming a dad (2.5-year-old stayed home), and I look forward to bringing my son to a CVPR one of these days! Big shoutout to all the moms, dads, and grandparents taking care of kids while the researchers mingle!

OrienterNet #CVPR2023 @pesarlin describing the advantages of an OrienterNet worldview, showing the demo, and jumping into details about Neural Map Matching!

#CVPR2023 Thursday panel discussion with non-CVPR-focused scientists: opportunistic forecasting, how a single number cannot communicate the value of a model, and bashing today’s SOTA-chasing culture in CV/ML. Lots of valuable insights!

OrienterNet: Visual Localization in 2D Public Maps with Neural Matching psarlin.com/orienternet If you love Neural Maps as much as us, come to Thursday evening poster 098 at #CVPR2023! We will be discussing our new approach to 3dof localization in 2D maps. #computervision

@pesarlin @Meta @RealityLabs @ddetone @shamangary @jstraub6 @rapideRobot Come say hi to us at the OrienterNet poster on this last day of CVPR! Today, Thursday, during the PM poster session which starts at 4:30pm. Poster 098.

OrienterNet uses global & free maps from OpenStreetMap so you can try the demo in *any* city in the world! Work done during an internship at @Meta @RealityLabs with @ddetone, @shamangary, A Avetisyan, @jstraub6, @quantombone, S Rota Bulo, @rapideRobot, P Kontschieder, V Balntas

📢 We just released the code for OrienterNet, a deep net that learns to localize in 2D maps 🗺 like OpenStreetMap, just like humans do! Join us today at #CVPR2023 poster PM-98. Video: youtu.be/wglW8jnupSs Code: github.com/facebookresear… Try our demo: take a picture and localize!

@BartronPolygon Hi Bart! The Project Aria glasses can continuously record for more than one hour. Take a look at our Project Aria wiki for hardware specs: facebookresearch.github.io/projectaria_to…

@quantombone Question since I'm unable to attend the conference this year: how long can these things continuously record for? Or operate at a lower sampling rate (e.g., 1 photo/sec, 1 photo / 5 seconds) for?

To learn about the sensor suite inside Project Aria glasses and see a demo of the data streaming to a computer for processing, come to the MetaAI booth at the #CVPR2023 expo

The brief discussion on “bringing your personality into research papers” was quite interesting. According to Dan, David Marr brought a lot of his personality into his work.

Dan Huttenlocher — “When I arrived at MIT for graduate school, I switched to Computer Vision because of David Marr and his theories” #CVPR2023

@LearnOpenCV It was a pleasure chatting with you tonight about OpenCV, Deep Learning, and open source! CVPR is special because it brings Twitter friends together under the same roof. #computervision

Carl Ren is talking about the software capabilities of Project Aria, the accompanying open source tools, and showcasing real-time capabilities. #CVPR23

Project Aria link to learn about becoming an Aria partner: projectaria.com/research-kit/

Jakob Engel talking about Project Aria Research Kit — what you get access to when you become an Aria partner. #CVPR23

Project Aria Tutorial: Richard Newcombe’s list of Computer Vision / Machine Perception Problems for this upcoming decade. #CVPR23

LightGlue from ETH Zurich presented at the image matching workshop. Seems to be a contender with SuperPoint+SuperGlue ! Supposedly, code release is coming soon #CVPR2023