Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2412.03907 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910894310555648 |
|---|---|
| author | Jin, Yizhou Zhu, Jiahui Wang, Guodong Li, Shiwei Zhang, Jinjin Liu, Xinyue Liu, Qingjie Wang, Yunhong |
| author_facet | Jin, Yizhou Zhu, Jiahui Wang, Guodong Li, Shiwei Zhang, Jinjin Liu, Xinyue Liu, Qingjie Wang, Yunhong |
| contents | Incremental anomaly detection aims to sequentially identify defects in industrial product lines but suffers from catastrophic forgetting, primarily due to knowledge overwriting during parameter updates and feature conflicts between tasks. In this work, We propose ONER (ONline Experience Replay), an end-to-end framework that addresses these issues by synergistically integrating two types of experience: (1) decomposed prompts, which dynamically generate image-conditioned prompts from reusable modules to retain prior knowledge thus prevent knowledge overwriting, and (2) semantic prototypes, which enforce separability in latent feature spaces at pixel and image levels to mitigate cross-task feature conflicts. Extensive experiments demonstrate the superiority of ONER, achieving state-of-the-art performance with +4.4% Pixel AUROC and +28.3% Pixel AUPR improvements on the MVTec AD dataset over prior methods. Remarkably, ONER achieves this with only 0.019M parameters and 5 training epochs per task, confirming its efficiency and stability for real-world industrial deployment. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_03907 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | ONER: Online Experience Replay for Incremental Anomaly Detection Jin, Yizhou Zhu, Jiahui Wang, Guodong Li, Shiwei Zhang, Jinjin Liu, Xinyue Liu, Qingjie Wang, Yunhong Computer Vision and Pattern Recognition Incremental anomaly detection aims to sequentially identify defects in industrial product lines but suffers from catastrophic forgetting, primarily due to knowledge overwriting during parameter updates and feature conflicts between tasks. In this work, We propose ONER (ONline Experience Replay), an end-to-end framework that addresses these issues by synergistically integrating two types of experience: (1) decomposed prompts, which dynamically generate image-conditioned prompts from reusable modules to retain prior knowledge thus prevent knowledge overwriting, and (2) semantic prototypes, which enforce separability in latent feature spaces at pixel and image levels to mitigate cross-task feature conflicts. Extensive experiments demonstrate the superiority of ONER, achieving state-of-the-art performance with +4.4% Pixel AUROC and +28.3% Pixel AUPR improvements on the MVTec AD dataset over prior methods. Remarkably, ONER achieves this with only 0.019M parameters and 5 training epochs per task, confirming its efficiency and stability for real-world industrial deployment. |
| title | ONER: Online Experience Replay for Incremental Anomaly Detection |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2412.03907 |