Context-encoding Variational Autoencoder for Unsupervised Anomaly Detection

12/14/2018
by   David Zimmerer, et al.
0

Unsupervised learning can leverage large-scale data sources without the need for annotations. In this context, deep learning-based auto encoders have shown great potential in detecting anomalies in medical images. However, state-of-the-art anomaly scores are still based on the reconstruction error, which lacks in two essential parts: it ignores the model-internal representation employed for reconstruction, and it lacks formal assertions and comparability between samples. We address these shortcomings by proposing the Context-encoding Variational Autoencoder (ceVAE) which combines reconstruction- with density-based anomaly scoring. This improves the sample- as well as pixel-wise results. In our experiments on the BraTS-2017 and ISLES-2015 segmentation benchmarks, the ceVAE achieves unsupervised ROC-AUCs of 0.95 and 0.89, respectively, thus outperforming state-of-the-art methods by a considerable margin.

READ FULL TEXT

page 9

page 10

research
11/28/2019

A Case for the Score: Identifying Image Anomalies using Variational Autoencoder Gradients

Through training on unlabeled data, anomaly detection has the potential ...
research
07/04/2019

Unsupervised Anomaly Localization using Variational Auto-Encoders

An assumption-free automatic check of medical images for potentially ove...
research
05/05/2020

Interpreting Rate-Distortion of Variational Autoencoder and Using Model Uncertainty for Anomaly Detection

Building a scalable machine learning system for unsupervised anomaly det...
research
12/12/2020

Anomaly detection through latent space restoration using vector-quantized variational autoencoders

We propose an out-of-distribution detection method that combines density...
research
11/26/2021

Data Invariants to Understand Unsupervised Out-of-Distribution Detection

Unsupervised out-of-distribution (U-OOD) detection has recently attracte...
research
05/19/2020

Unsupervised anomaly localization using VAE and beta-VAE

Variational Auto-Encoders (VAEs) have shown great potential in the unsup...
research
09/06/2020

Deep Learning for the Analysis of Disruption Precursors based on Plasma Tomography

The JET baseline scenario is being developed to achieve high fusion perf...

Please sign up or login with your details

Forgot password? Click here to reset