Research map / 3D vision / Gaussian splatting
IQGS: Instance Query-based Gaussian Segmentation
Yichao Gao, Xinyuan Liu, Yike Ma, Yucheng Zhang, Feng Dai
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 focal loss (ablation) | IQGS · LERF | mIoU (%): 33.91 -> 59.36 | Improved |
| add dice loss (ablation) | IQGS · LERF | mIoU (%): 49.43 -> 59.36 | Improved |
| add 3D regularization loss (ablation) | IQGS · 360 v2 | mIoU (%): 60.08 -> 72.02 | Improved |
| add 2D-3D consistency loss (ablation) | IQGS · 360 v2 | mIoU (%): 66.17 -> 72.02 | Improved |
| use ID matching instead of Hungarian matching (ablation) | IQGS · 360 v2 | mIoU (%): 58.21 -> 72.02 | Improved |
| supervise with binary masks instead of ID maps (ablation) | IQGS · LERF | mIoU (%): 40.02 -> 59.36 | Improved |
| increase query number Nq from 50 to 300 (D=16) (ablation) | IQGS · 360 v2 | mIoU (%): 57.13 -> 72.02 | Improved |
| feature dimension D=16 instead of 8 or 32 (Nq=300) (ablation) | IQGS · LERF | mIoU (%): 52.68 -> 59.36 | Improved |
| lower lambda_mask from 20.0 to 10.0 (ablation) | IQGS · LERF | mIoU (%): 59.36 -> 51.15 | Hurt |
| raise lambda_dice from 1.0 to 2.0 (ablation) | IQGS · LERF | mIoU (%): 59.36 -> 56.52 | Hurt |
| lower lambda_reg from 1.0 to 0.5 (ablation) | IQGS · 360 v2 | mIoU (%): 72.02 -> 70.34 | Hurt |
| apply Gaussian Grouping to 2DGS instead of 3DGS | Gaussian Grouping · mean mIoU over scenes | mIoU (%): 57.17 -> 36.34 | Hurt |
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.
- Gaussian Grouping — Gaussian Grouping: Segment and Edit Anything in 3D Scenes (ECCV 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
- Bootstraping Clustering of Gaussians for View-consistent 3D Scene Understanding (AAAI 2025 · similarity 0.87)
- SegVGGT: Joint 3D Reconstruction and Instance Segmentation from Multi-View Images (ECCV 2026 · similarity 0.86)
- B$^3$-Seg: Camera-Free, Training-Free 3DGS Segmentation via Analytic EIG and Beta-Bernoulli Bayesian Updates (CVPR 2026 · similarity 0.86)
- PanoGS: Gaussian-based Panoptic Segmentation for 3D Open Vocabulary Scene Understanding (CVPR 2025 · similarity 0.85)
- GaussianCut: Interactive segmentation via graph cut for 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.85)
- Large Images Are Gaussians: High-Quality Large Image Representation with Levels of 2D Gaussian Splatting (AAAI 2025 · similarity 0.85)
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.
