I dig the unity that is happening across Canada. Coast to coast. First nations, provinces, and people from all walks of life. I hope we can built from that.
Wow. Recreating the Shawshank Redemption prison in 3D from a single video, in real time (!)
Just read the MASt3R-SLAM paper and it's pretty neat. These folks basically built a real-time dense SLAM system on top of MASt3R, which is a transformer-based neural network that can do 3d reconstruction and localization from uncalibrated image pairs.
The cool part is they don't need a fixed camera model -- it just works with arbitrary cameras -- think different focal lengths, sensor sizes, even handling zooming in video (FMV drone video anyone?!). If you've done photogrammetry or played with NeRFs you know that is a HUGE deal.
They've solved some tricky problems like efficient point matching and tracking, plus they've figured out how to fuse point clouds and handle loop closures in real-time.
Their system runs at about 15 FPS on a 4090 and produces both camera poses and dense geometry. When they know the camera calibration, they get SOTA results across several benchmarks, but even without calibration, they still perform well.
What's interesting is the approach -- most recent SLAM work has built on DROID-SLAM's architecture, but these folks went a different direction by leveraging a strong 3D reconstruction prior. Seems to give them more coherent geometry, which makes sense since that's what MASt3R was designed for.
For anyone who cares about monocular SLAM and 3D reconstruction, this feels like a significant step toward plug-and-play dense SLAM without calibration headaches -- perfect for drones, robots, AR/VR -- the works!
@EHuanglu I think this is smoke and mirrors. I'm playing around with it....most if not all the geo created has holes and bad geo. I'm guesing they might have started with a premade terrible asset, then got a 3d artist to clean it up and repaint.
What's up with the memes? Here is an update from Francesco on the Blender Development Fund donation campaign donation campaign. #b3d#devfund#jointhetwopercent https://t.co/Tf18irBnrE
Spent the last few weeks parsing NYC's massive 2017 lidar dataset (almost 1TB of data) and segmenting them by building. Now you can select any of NYC's more than 1M buildings and see its point cloud. #webgpu
The way this video started, I thought it was going to be some comedy sketch, but this is actually something that could save me loads of time when it comes to mural work.
via https://t.co/UjgXkF2Z1E