Semi-supervised Anomaly Detection using AutoEncoders

TL;DR

This paper introduces a convolutional auto-encoder method for anomaly detection that trains solely on normal data, effectively identifying defects in images without using defect examples during training.

Contribution

It presents a novel semi-supervised auto-encoder architecture that detects anomalies by learning only from defect-free images, enabling effective defect segmentation.

Findings

Achieved an average F1 score of 0.885 on two datasets.

Successfully detected defect shapes without training on defected images.

Demonstrated robustness in industrial defect detection scenarios.

Abstract

Anomaly detection refers to the task of finding unusual instances that stand out from the normal data. In several applications, these outliers or anomalous instances are of greater interest compared to the normal ones. Specifically in the case of industrial optical inspection and infrastructure asset management, finding these defects (anomalous regions) is of extreme importance. Traditionally and even today this process has been carried out manually. Humans rely on the saliency of the defects in comparison to the normal texture to detect the defects. However, manual inspection is slow, tedious, subjective and susceptible to human biases. Therefore, the automation of defect detection is desirable. But for defect detection lack of availability of a large number of anomalous instances and labelled data is a problem. In this paper, we present a convolutional auto-encoder architecture for anomaly detection that is trained only on the defect-free (normal) instances. For the test images, residual masks that are obtained by subtracting the original image from the auto-encoder output are thresholded to obtain the defect segmentation masks. The approach was tested on two data-sets and achieved an impressive average F1 score of 0.885. The network learnt to detect the actual shape of the defects even though no defected images were used during the training.