iiinova — where your digital self keeps evolving

Bart Duisterhof

@bardienus.bsky.social

2.3K followers 691 following 28 posts

PhD Student @cmurobotics.bsky.social with @jeff-ichnowski.bsky.social || DUSt3R Research Intern @naverlabseurope || 4D Vision for Robot Manipulation 📷 He/Him - bart-ai.com

Posts Replies Media

Bart Duisterhof @bardienus.bsky.social · 19/09/2025

RaySt3R was accepted to NeurIPS! Check out the HuggingFace demo for image to 3D in cluttered scenes huggingface.co/spaces/bartd...

052

Bart Duisterhof @bardienus.bsky.social · 06/06/2025

We also study the impact of the confidence threshold on reconstruction quality. Our ablations suggest setting a higher confidence threshold improves accuracy, while limiting completeness and edge-bleeding. Users can tune the threshold for application-specific requirements 🎛️.

100

Bart Duisterhof @bardienus.bsky.social · 06/06/2025

We evaluate RaySt3R against the baselines on synthetic and real-world datasets. The results suggest RaySt3R achieves zero-shot generalization to the real world, and outperforms all baselines by up to 44% in 3D chamfer distance 🚀.

100

Bart Duisterhof @bardienus.bsky.social · 06/06/2025

We train RaySt3R by curating a new dataset, for a total of 12 million views 📷 with Objaverse and GSO objects. The ablations 🔍 suggest that more and more diverse data improves RaySt3R's performance. RaySt3R does not require GT meshes, paving the way for training on real-world data.

100

Bart Duisterhof @bardienus.bsky.social · 06/06/2025

💡 Our key insight is that 3D object shape completion can be recasted as a novel-view synthesis problem. RaySt3R takes a masked RGB-D image as input, and predicts depth maps and object masks for novel views. We query multiple views and merge the predictions into a consistent point cloud.

110

Bart Duisterhof @bardienus.bsky.social · 06/06/2025

Imagine if robots could fill in the blanks in cluttered scenes. ✨ Enter RaySt3R: a single masked RGB-D image in, complete 3D out. It infers depth, object masks, and confidence for novel views, and merges the predictions into a single point cloud. rayst3r.github.io

1243

Bart Duisterhof @bardienus.bsky.social · 18/11/2024

Hello world! I'm a PhD student at CMU in robotics, at the intersection of vision and robot manipulation.

1313