Skip to content
MeetupOnlinePast event

Best of 3DV 2026: Live Streams from the 3D Vision Conference

Event artwork for Best of 3DV 2026: Live Streams from the 3D Vision Conference

This event has ended

This listing is kept as an archive. Explore upcoming events to find your next one.

See upcoming events
About this event

What to expect

Experience cutting-edge 3D vision research from the comfort of your home with this virtual showcase of the best presentations from the 3DV 2026 conference. This online meetup brings together AI and computer vision enthusiasts to watch live streams of select conference talks, featuring groundbreaking work in 3D scene reconstruction, dynamic content generation, and few-shot object perception.

Featured Presentations

The session opens with Navigating a 3D Vision Conference with VLMs and Embeddings, where Harpreet Sahota demonstrates how to systematically map an entire academic conference using modern AI tools. Learn how document embeddings from Jina v4 and vision-language models can transform 177 research papers into an interactive, queryable dataset — a practical approach to knowledge discovery that goes far beyond traditional title-skimming.

Rio Aguina-Kang from Drafted AI presents Seeing Through Clutter, introducing an innovative iterative pipeline that reconstructs structured 3D scenes from single images. By removing foreground objects one at a time using VLMs as orchestrators, this method achieves state-of-the-art robustness in highly occluded environments — no task-specific training required.

Lu Sang from TU Munich delivers Physical Realistic 4D Generation, covering a series of works that bring physical realism to dynamic 3D content. From neural surface deformation to the TwoSquared framework (3DV 2026 oral presentation), discover how implicit neural representations with physically constrained velocity fields enable controllable 4D content creation from just two 2D image pairs.

The session concludes with Sebastian Jung from the German Aerospace Center presenting Finding NeMO, a geometry-aware representation that enables 6DoF object pose estimation, detection, and segmentation from only a handful of RGB images — solving the challenge of perceiving unknown objects without retraining models.

Who Should Attend

This event is ideal for machine learning engineers, computer vision researchers, AI practitioners, and anyone interested in 3D reconstruction, vision-language models, or the latest academic research in spatial computing. Whether you're building AR/VR applications, working on robotics perception, or simply curious about state-of-the-art computer vision techniques, you'll gain valuable insights from these expert presentations.

The virtual format makes this accessible to a global audience, with speakers presenting live and op

Running an AI event?

List it here, free.

Meetups, workshops and conferences go in front of thousands of builders after a quick review.

List your event