Research map / Label-efficient and robust learning / OOD and anomaly detection
Complementary Prototype Mapping for Efficient Multimodal Anomaly Detection
Yuan Zhao, Xiaoqin Zhang, Huchuan Lu, Lihe Zhang
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 consensus prototypes to the MLP mapping baseline (ablation) | CPMAD baseline · MVTec-3D | I-AUROC: 89.5 -> 95.9 | Improved |
| add supplementary prototypes on top of consensus prototypes (ablation) | CPMAD · MVTec-3D | I-AUROC: 95.9 -> 97.8 | Improved |
| use CR-Attention instead of standard cross-attention in SQM (ablation) | CPMAD · MVTec-3D | I-AUROC: 96.8 -> 97.8 | Improved |
| use non-learnable average instead of learnable weighted average in CEM (ablation) | CPMAD · MVTec-3D | I-AUROC: 97.3 -> 97.8 | Improved |
| use dynamic anchors instead of static learnable anchors (ablation) | CPMAD · MVTec-3D | I-AUROC: 97.2 -> 97.8 | Improved |
| use channel-wise residual instead of token-wise residual in mapping (ablation) | CPMAD · MVTec-3D | I-AUROC: 97.6 -> 97.8 | Improved |
| use CPMAD instead of CFM in 5-shot setting | few-shot MAD · MVTec-3D 5-shot | I-AUROC: 81.1 -> 89.9 | Improved |
| use CPMAD instead of CFM in 5-shot setting | few-shot MAD · Eyecandies 5-shot | I-AUROC: 79.5 -> 88.0 | Improved |
| use lightweight ViT-S variant CPMAD-S instead of CPMAD | CPMAD · MVTec-3D | I-AUROC: 97.8 -> 96.6 | Mixed |
| increase SQM query size / CEM anchor size beyond a moderate value (ablation) | CPMAD · MVTec-3D | I-AUROC | No clear effect |
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.
- G2SF — G2SF: Geometry-Guided Score Fusion for Multimodal Industrial Anomaly Detection (ICCV 2025) · better on 1 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
- PIRN: Prototypical-based Intra-modal Reconstruction with Normality Communication for Multi-modal Anomaly Detection. (ICLR 2026 · similarity 0.90)
- Beyond Single-Modal Boundary: Cross-Modal Anomaly Detection through Visual Prototype and Harmonization (CVPR 2025 · similarity 0.88)
- Multimodal Industrial Anomaly Detection by Crossmodal Feature Mapping (CVPR 2024 · similarity 0.87)
- Rethinking Reverse Distillation for Multi-Modal Anomaly Detection (AAAI 2024 · similarity 0.85)
- Supervision Recovery for Time Series Anomaly Detection via Counterfactual Pairing (NeurIPS 2026 · similarity 0.85)
- Towards Training-free Anomaly Detection with Vision and Language Foundation Models (CVPR 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.
