SceneFun3D: 3D Indoor Scenes with Functional Annotations and Multimodal Captures

SceneFun3D is a 3D scene-understanding dataset of high-resolution Faro laser-scan point clouds of indoor environments. The dataset is densely annotated with fine-grained functional interactive elements, their affordances, motion parameters, and free-form task descriptions. Each scene is also captured by several iPad video sequences with RGB, depth, camera poses, and intrinsics.

Use Cases

• Train models for 3D object affordance recognition based on annotated functional elements like handles and switches.
• Develop algorithms for robotic manipulation planning using motion parameters and task descriptions.
• Research multimodal scene representation learning by aligning point clouds with RGB-D video sequences.
• Benchmark 3D semantic segmentation and instance detection in complex indoor environments.

Strengths

• Dataset includes high-resolution Faro laser-scan point clouds, suggesting precise 3D geometry.
• Annotations include fine-grained functional elements, affordances, motion parameters, and free-form task descriptions, indicating rich semantic labels.
• Multimodal data capture includes several iPad video sequences per scene with RGB, depth, camera poses, and intrinsics.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last updated 2026-06-29 19:44:43; freshness should be verified.

Provenance

Source

Voxel51

Collection Method

Likely collected via Faro laser scanning and iPad video capture in indoor environments.

Freshness

Last updated 2026-06-29 19:44:43.