Research map / Point clouds / Point cloud classification
Back to Point: Exploring Point-Language Models for Zero-Shot 3D Anomaly Detection
Kaiqiang Li, Gang Li, Mingle Zhou, Min Li, Delong Han, Jin Wan
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 MGFEM to baseline (ablation) | BTP baseline · Real3D-AD | point-level AUROC: - -> 83.3 | Improved |
| add MGFEM to baseline (ablation) | BTP baseline · Real3D-AD | O-AUROC: 51.5 -> 59.9 | Improved |
| add GFCM alone to baseline (ablation) | BTP baseline · Real3D-AD | O-AUROC: 51.5 -> 52.5 | No clear effect |
| add GFCM on top of MGFEM (ablation) | BTP · Real3D-AD | O-AUROC: 59.9 -> 61.4 | Improved |
| add GFCM on top of MGFEM (ablation) | BTP · Anomaly-ShapeNet | point-level AUROC: 83.5 -> 87.3 | Improved |
| add geometric loss (ablation) | BTP · Anomaly-ShapeNet | point-level AUROC: 74.1 -> 87.3 | Improved |
| add geometric loss (ablation) | BTP · Real3D-AD | point-level AUROC: 82.2 -> 84.5 | Improved |
| add local loss (ablation) | BTP · Real3D-AD | point-level AUROC: 68.8 -> 84.5 | Improved |
| add global loss (ablation) | BTP · Real3D-AD | O-AUROC: 60.5 -> 61.4 | Improved |
| add global loss (ablation) | BTP · Real3D-AD | point-level PRO: 59.1 -> 55.5 | Hurt |
| use 1024 instead of 2048 input points (ablation) | BTP · Real3D-AD | P-AUROC: 83.0 -> 84.5 | Hurt |
| use 8192 instead of 2048 input points (ablation) | BTP · Real3D-AD | P-AUROC: 84.5 -> 80.1 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 3 on most shared results.
- PointAD — PointAD: Comprehending 3D Anomalies from Points and Pixels for Zero-shot 3D Anomaly Detection (NeurIPS 2024) · better on 1 of 3 shared results
- PointCLIPV2 — PointCLIP V2: Prompting CLIP and GPT for Powerful 3D Open-world Learning (ICCV 2023) · better on 2 of 2 shared results
- R3DAD — R3D-AD: Reconstruction via Diffusion for 3D Anomaly Detection (ECCV 2024) · better on 0 of 1 shared result
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
- PointAD: Comprehending 3D Anomalies from Points and Pixels for Zero-shot 3D Anomaly Detection (NeurIPS 2024 · similarity 0.91)
- Towards Scalable 3D Anomaly Detection and Localization: A Benchmark via 3D Anomaly Synthesis and A Self-Supervised Learning Network (CVPR 2024 · similarity 0.90)
- Towards Generalizable 3D Anomaly Detection via Relational Inconsistency Modeling (NeurIPS 2026 · similarity 0.88)
- Registration is a Powerful Rotation-Invariance Learner for 3D Anomaly Detection (NeurIPS 2025 · similarity 0.88)
- Looking 3D: Anomaly Detection with 2D-3D Alignment (CVPR 2024 · similarity 0.88)
- PO3AD: Predicting Point Offsets toward Better 3D Point Cloud Anomaly Detection (CVPR 2025 · similarity 0.87)
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.
