Skip to content
MeetupOnlinePast event

Best of 3DV 2026: Live-Streamed Computer Vision Research

Event artwork for Best of 3DV 2026: Live-Streamed Computer Vision Research

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

Best of 3DV 2026 brings the year's most groundbreaking 3D vision research directly to your screen through a live-streamed online event on May 11, 2026. Organized by the SF Bay Area AI Meetup Group, this virtual conference digest showcases cutting-edge advances in 3D computer vision, from physically realistic 4D generation to few-shot object perception and scene reconstruction.

What You'll Learn

This event features four research presentations covering the latest developments in 3D vision and AI:

Navigating Academic Conferences with AI — Harpreet Sahota, hacker-in-residence and ML engineer, demonstrates how to process and visualize all 177 papers from 3DV 2026 using vision-language models and embeddings. Learn practical techniques for building interactive conference maps using open-source tools, document embeddings from Jina v4, and FiftyOne Brain for UMAP visualization and similarity search.

Structured 3D Scene Reconstruction — Rio Aguina-Kang from Drafted AI presents SeeingThroughClutter, a novel method that reconstructs structured 3D representations from single images by iteratively removing and modeling objects. This approach leverages VLMs as orchestrators to achieve state-of-the-art robustness on 3D-Front and ADE20K datasets, particularly in highly occluded scenes.

Physical Realistic 4D Generation — Lu Sang, PhD researcher at TU Munich, shares a series of works on generating dynamic 3D content that moves and deforms over time. Her presentation covers neural surface deformation with velocity fields (ICLR 2025), the 4Deform framework (CVPR 2025), and TwoSquared (3DV 2026 oral) — which achieves full 4D generation from just two 2D image pairs.

Few-Shot Object Perception — Sebastian Jung, PhD candidate at the German Aerospace Center (DLR) and Google student researcher, introduces NeMO, a geometry-aware representation that enables 6DoF object pose estimation, detection, and segmentation using only a handful of RGB images — no model retraining required.

Who Should Attend

This event is ideal for computer vision researchers, machine learning engineers, and AI practitioners interested in 3D reconstruction, VLMs, document AI, robotics applications, and generative visual content. Whether you're building XR experiences, working on physic

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