Exploring normalizing flow for anomaly detection

Master thesis (2019)

Authors

C.A. Pathak Electrical Engineering, Mathematics and Computer Science

Contributors

J.C. van Gemert Pattern Recognition and Bioinformatics - EEMCS (supervisor 1)
M.J.T. Reinders Pattern Recognition and Bioinformatics - EEMCS (supervisor 2)
A.J. van Genderen Computer Engineering - EEMCS (supervisor 2)
Faculty
Electrical Engineering, Mathematics and Computer Science
Copyright: © 2019 Chinmay Pathak
Autoencoder Anomaly Detection Outlier detection Generative Algorithms Unsupervised learning One-class classification GLOW Normalizing flows
More Info
expand_more

Published Date

29-08-2019

Language

English

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Faculty

Electrical Engineering, Mathematics and Computer Science

Abstract

Anomaly detection is a task of interest in many domains. Typical way of tackling this problem is using an unsupervised way. Recently, deep neural network based density estimators such as Normalizing flows have seen a huge interest. The ability of these models to do the exact latent-variable inference and exact log-likelihood calculation with invertible architecture makes them interesting for the task of anomaly detection. In this work we explore the such normalizing flow-based model approach for anomaly detection in the novel BoroscopeV1 dataset which contains videos of the actual industry boroscope video material and has large noise. We verify the correctness of the models on a toy dataset. We found that the black pixels and high frequency in the image affect the model likelihood adversely. The experimental evidence shows that the normalizing flow-based approach can be used for the task of anomaly detection