MECAD: A multi-expert architecture for continual anomaly detection
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866909967380905984 |
|---|---|
| author | Dahmardeh, Malihe Setti, Francesco |
| author_facet | Dahmardeh, Malihe Setti, Francesco |
| contents | In this paper we propose MECAD, a novel approach for continual anomaly detection using a multi-expert architecture. Our system dynamically assigns experts to object classes based on feature similarity and employs efficient memory management to preserve the knowledge of previously seen classes. By leveraging an optimized coreset selection and a specialized replay buffer mechanism, we enable incremental learning without requiring full model retraining. Our experimental evaluation on the MVTec AD dataset demonstrates that the optimal 5-expert configuration achieves an average AUROC of 0.8259 across 15 diverse object categories while significantly reducing knowledge degradation compared to single-expert approaches. This framework balances computational efficiency, specialized knowledge retention, and adaptability, making it well-suited for industrial environments with evolving product types. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_15323 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MECAD: A multi-expert architecture for continual anomaly detection Dahmardeh, Malihe Setti, Francesco Computer Vision and Pattern Recognition In this paper we propose MECAD, a novel approach for continual anomaly detection using a multi-expert architecture. Our system dynamically assigns experts to object classes based on feature similarity and employs efficient memory management to preserve the knowledge of previously seen classes. By leveraging an optimized coreset selection and a specialized replay buffer mechanism, we enable incremental learning without requiring full model retraining. Our experimental evaluation on the MVTec AD dataset demonstrates that the optimal 5-expert configuration achieves an average AUROC of 0.8259 across 15 diverse object categories while significantly reducing knowledge degradation compared to single-expert approaches. This framework balances computational efficiency, specialized knowledge retention, and adaptability, making it well-suited for industrial environments with evolving product types. |
| title | MECAD: A multi-expert architecture for continual anomaly detection |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.15323 |