Outliers resistant image classification by anomaly detection

TL;DR

This paper introduces a model combining classification and anomaly detection using metric learning to improve robustness of image classification in production monitoring, especially against outliers and environmental variations.

Contribution

It proposes a novel approach that integrates anomaly detection with classification via metric learning, addressing the challenge of unpredictable outliers in production environments.

Findings

Effective in detecting outliers and environmental variations

Compared multiple model architectures with promising results

Enhanced robustness of image classification in manual assembly monitoring

Abstract

Various technologies, including computer vision models, are employed for the automatic monitoring of manual assembly processes in production. These models detect and classify events such as the presence of components in an assembly area or the connection of components. A major challenge with detection and classification algorithms is their susceptibility to variations in environmental conditions and unpredictable behavior when processing objects that are not included in the training dataset. As it is impractical to add all possible subjects in the training sample, an alternative solution is necessary. This study proposes a model that simultaneously performs classification and anomaly detection, employing metric learning to generate vector representations of images in a multidimensional space, followed by classification using cross-entropy. For experimentation, a dataset of over 327,000 images was prepared. Experiments were conducted with various computer vision model architectures, and the outcomes of each approach were compared.