Multi-Level Feature Fusion for Continual Learning in Visual Quality Inspection

TL;DR

This paper introduces a multi-level feature fusion method for continual learning in visual quality inspection, enabling efficient adaptation to new conditions while minimizing forgetting and maintaining high performance.

Contribution

The proposed MLFF approach leverages features from different network depths to enhance continual learning efficiency and robustness in manufacturing inspection tasks.

Findings

Matches end-to-end training performance with fewer parameters

Reduces catastrophic forgetting effectively

Improves generalization to new defect types

Abstract

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.