Multi-Level Feature Fusion for Continual Learning in Visual Quality Inspection
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908827254784000 |
|---|---|
| author | Bauer, Johannes C. Geng, Paul Trattnig, Stephan Dokládal, Petr Daub, Rüdiger |
| author_facet | Bauer, Johannes C. Geng, Paul Trattnig, Stephan Dokládal, Petr Daub, Rüdiger |
| contents | Deep neural networks show great potential for automating various visual quality inspection tasks in manufacturing. However, their applicability is limited in more volatile scenarios, such as remanufacturing, where the inspected products and defect patterns often change. In such settings, deployed models require frequent adaptation to novel conditions, effectively posing a continual learning problem. To enable quick adaptation, the necessary training processes must be computationally efficient while still avoiding effects like catastrophic forgetting. This work presents a multi-level feature fusion (MLFF) approach that aims to improve both aspects simultaneously by utilizing representations from different depths of a pretrained network. We show that our approach is able to match the performance of end-to-end training for different quality inspection problems while using significantly less trainable parameters. Furthermore, it reduces catastrophic forgetting and improves generalization robustness to new product types or defects. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_00725 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Multi-Level Feature Fusion for Continual Learning in Visual Quality Inspection Bauer, Johannes C. Geng, Paul Trattnig, Stephan Dokládal, Petr Daub, Rüdiger Computer Vision and Pattern Recognition Deep neural networks show great potential for automating various visual quality inspection tasks in manufacturing. However, their applicability is limited in more volatile scenarios, such as remanufacturing, where the inspected products and defect patterns often change. In such settings, deployed models require frequent adaptation to novel conditions, effectively posing a continual learning problem. To enable quick adaptation, the necessary training processes must be computationally efficient while still avoiding effects like catastrophic forgetting. This work presents a multi-level feature fusion (MLFF) approach that aims to improve both aspects simultaneously by utilizing representations from different depths of a pretrained network. We show that our approach is able to match the performance of end-to-end training for different quality inspection problems while using significantly less trainable parameters. Furthermore, it reduces catastrophic forgetting and improves generalization robustness to new product types or defects. |
| title | Multi-Level Feature Fusion for Continual Learning in Visual Quality Inspection |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2601.00725 |