Research map / 3D vision / 3D generation
ASSIST-3D: Adapted Scene Synthesis for Class-Agnostic 3D Instance Segmentation
Shengchao Zhou, Jiehong Lin, Jiahui Liu, Shizhen Zhao, Chirui Chang, Xiaojuan Qi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add ASSIST-3D synthetic data to ScanNetV2 training | Mask3D · ScanNet++ | AP: 12.0 -> 22.2 | Improved |
| add ASSIST-3D synthetic data to ScanNetV2 training | Mask3D · S3DIS | AP: 13.6 -> 29.0 | Improved |
| add ASSIST-3D synthetic data to ScanNetV2 training | Mask3D · ScanNetV2 val | AP: 46.6 -> 48.1 | Improved |
| increase number of object geometry clusters used for synthesis from 1 to 5 (ablation) | ASSIST-3D · ScanNet++ | AP: 14.6 -> 22.2 | Improved |
| increase geometry clusters from 1 to 5 (ablation) | ASSIST-3D · ScanNet++ (out-of-domain objects) | AR: 3.2 -> 12.4 | Improved |
| lower co-occurrence enforcement (100% to 0%) to raise context complexity (ablation) | ASSIST-3D · ScanNet++ | AP: 17.2 -> 22.2 | Improved |
| use rendering-based point cloud construction instead of mesh sampling (ablation) | ASSIST-3D · S3DIS | AP: 18.8 -> 29.0 | Improved |
| use ASSIST-3D scene layout/selection instead of Holodeck data (ablation) | Mask3D · ScanNet++ | AP: 14.2 -> 22.2 | Improved |
| use reasonable layouts (ASSIST-3D) instead of RandomRooms random placement (ablation) | Mask3D · S3DIS | AP: 23.5 -> 29.0 | Improved |
| scale number of synthetic scenes and object classes (ablation) | ASSIST-3D · ScanNet++ and S3DIS | AP | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SA3DIP — SA3DIP: Segment Any 3D Instance with Potential 3D Priors (NeurIPS 2024) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- SAI3D: Segment Any Instance in 3D Scenes (CVPR 2024 · similarity 0.88)
- Open3DIS: Open-Vocabulary 3D Instance Segmentation with 2D Mask Guidance (CVPR 2024 · similarity 0.88)
- Any3DIS: Class-Agnostic 3D Instance Segmentation by 2D Mask Tracking (CVPR 2025 · similarity 0.87)
- Segment Any 3D Object with Language (ICLR 2025 · similarity 0.87)
- PartCrafter: Structured 3D Mesh Generation via Compositional Latent Diffusion Transformers (NeurIPS 2025 · similarity 0.86)
- GrabS: Generative Embodied Agent for 3D Object Segmentation without Scene Supervision (ICLR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
