Proceedings Track
14-page full papers for inclusion in the ECCV workshop proceedings; subject to dual-submission policy, must be original work.
Computer Vision/CORE Unranked
CVPR 2026
CVPR 2026 Workshop on Open-World 3D Scene Understanding with Foundation Models (OpenSUN3D)
The OpenSUN3D 2026 workshop, held in conjunction with ECCV 2026 in Malmö, Sweden, focuses on open-world 3D scene understanding and representations, bringing together researchers to advance spatial intelligence, 3D perception, and generative world models for intelligent agents. The workshop features a challenge on 3D functionality, affordance, and interaction understanding, along with paper submissions in archival and non-archival tracks, including a new Huawei-sponsored track on memory-centric spatial intelligence.
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Non-Archival Track
7-page extended abstracts or 14-page full papers not included in the proceedings; may include previously published work.
Huawei SpatialMem Track
A dedicated sub-track for memory-centric, online, dynamic/4D or world-model-coupled spatial intelligence; accepts both archival and non-archival submissions, flagged during submission.
Research areas in scope
01
Topics of Interest
Dynamic & Hierarchical Spatial Representations — persistent open-vocabulary 3D scene graphs (metric, semantic, topological, 4D) and multi-view-consistent, metric-grounded VLM reasoningTest-Time Memory Evolution & Updating — updating, editing and bounding a 3D memory at inference across revisits; consolidation, selective forgetting, and memory for video world modelsNeuro-Symbolic Spatial Reasoning — verifiable programs, tools and graph-query interfaces over 3D scene graphs for checkable inferenceSystems & Efficiency — on-device, power/memory-budgeted mapping + reasoning; constant-memory streaming maps and deployment-overhead benchmarksMulti-Session Persistent Spatial Memory — representations and benchmarks for memory persisting across sessions (hours to days+), tracking identity, relations, uncertainty and provenance
02
Challenge Tracks
Functionality Segmentation [SceneFun3D]Open-Vocabulary 3D Affordance Grounding [SceneFun3D]Interaction Understanding [Articulate3D]
Policies worth checking twice
- All submissions must be anonymous.
- Submissions must follow the ECCV 2026 author guidelines.
- Proceedings-track papers are subject to dual-submission rules (cannot be under review or published elsewhere).
- Non-archival submissions may include previously published work and are exempt from dual-submission restrictions.
- Authors may submit to the main track and also flag their paper as SpatialMem for consideration in the sponsored sub-track.
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026