What do we learn? Debunking the Myth of Unsupervised Outlier Detection

by   Cosmin I. Bercea, et al.

Even though auto-encoders (AEs) have the desirable property of learning compact representations without labels and have been widely applied to out-of-distribution (OoD) detection, they are generally still poorly understood and are used incorrectly in detecting outliers where the normal and abnormal distributions are strongly overlapping. In general, the learned manifold is assumed to contain key information that is only important for describing samples within the training distribution, and that the reconstruction of outliers leads to high residual errors. However, recent work suggests that AEs are likely to be even better at reconstructing some types of OoD samples. In this work, we challenge this assumption and investigate what auto-encoders actually learn when they are posed to solve two different tasks. First, we propose two metrics based on the Fréchet inception distance (FID) and confidence scores of a trained classifier to assess whether AEs can learn the training distribution and reliably recognize samples from other domains. Second, we investigate whether AEs are able to synthesize normal images from samples with abnormal regions, on a more challenging lung pathology detection task. We have found that state-of-the-art (SOTA) AEs are either unable to constrain the latent manifold and allow reconstruction of abnormal patterns, or they are failing to accurately restore the inputs from their latent distribution, resulting in blurred or misaligned reconstructions. We propose novel deformable auto-encoders (MorphAEus) to learn perceptually aware global image priors and locally adapt their morphometry based on estimated dense deformation fields. We demonstrate superior performance over unsupervised methods in detecting OoD and pathology.


page 2

page 4

page 5

page 7

page 8

page 14


Anomaly detection in image or latent space of patch-based auto-encoders for industrial image analysis

We study several methods for detecting anomalies in color images, constr...

A Scene-Agnostic Framework with Adversarial Training for Abnormal Event Detection in Video

Abnormal event detection in video is a complex computer vision problem t...

Unsupervised Abnormality Detection through Mixed Structure Regularization (MSR) in Deep Sparse Autoencoders

Deep sparse auto-encoders with mixed structure regularization (MSR) in a...

Object-centric Auto-encoders and Dummy Anomalies for Abnormal Event Detection in Video

Abnormal event detection in video is a challenging vision problem. Most ...

Unsupervised heart abnormality detection based on phonocardiogram analysis with Beta Variational Auto-Encoders

Heart Sound (also known as phonocardiogram (PCG)) analysis is a popular ...

Learning Correlated Latent Representations with Adaptive Priors

Variational Auto-Encoders (VAEs) have been widely applied for learning c...

Localisation of Mammographic masses by Greedy Backtracking of Activations in the Stacked Auto-Encoders

Mammographic image analysis requires accurate localisation of salient ma...

Please sign up or login with your details

Forgot password? Click here to reset