Exploring Deep Anomaly Detection Methods Based on Capsule Net
This work addresses anomaly detection for image data, but it is incremental as it adapts existing capsule network techniques to this domain.
The paper tackled anomaly detection in image data by developing methods based on capsule networks, resulting in CapsNet-based approaches that outperformed benchmark methods in many cases, with the prediction-probability-based method performing consistently well.
In this paper, we develop and explore deep anomaly detection techniques based on the capsule network (CapsNet) for image data. Being able to encoding intrinsic spatial relationship between parts and a whole, CapsNet has been applied as both a classifier and deep autoencoder. This inspires us to design a prediction-probability-based and a reconstruction-error-based normality score functions for evaluating the "outlierness" of unseen images. Our results on three datasets demonstrate that the prediction-probability-based method performs consistently well, while the reconstruction-error-based approach is relatively sensitive to the similarity between labeled and unlabeled images. Furthermore, both of the CapsNet-based methods outperform the principled benchmark methods in many cases.