Unsupervised Anomaly Detection with Generative Adversarial Networks to Guide Marker Discovery

By Thomas Schlegl, Philipp Seeböck, Sebastian M. Waldstein, Ursula Schmidt-Erfurth, Georg Langs. 2017

2,527 citations recorded by the source corpus.

Obtaining models that capture imaging markers relevant for disease progression and treatment monitoring is challenging. Models are typically based on large amounts of data with annotated examples of known markers aiming at automating detection. High annotation effort and the limitation to a vocabulary of known markers limit the power of such approaches. Here, we perform unsupervised learning to identify anomalies in imaging data as candidates for markers. We propose AnoGAN, a deep convolutional generative adversarial network to learn a manifold of normal anatomical variability, accompanying a novel anomaly scoring scheme based on the mapping from image space to a latent space. Applied to new data, the model labels anomalies, and scores image patches indicating their fit into the learned distribution. Results on optical coherence tomography images of the retina demonstrate that the approach correctly identifies anomalous images, such as images containing retinal fluid or hyperreflective foci.

Subfields: Artificial Intelligence, Biophysics, Radiology, Nuclear Medicine and Imaging

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Biophysics

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Recent Progress on Generative Adversarial Networks (GANs): A Survey (2019, background): “In medicine, Schlegl et al. [65] proposed an AnoGAN for anomaly detection of medical images, and learned the characteristics of lesions by learning the characteristics of health data sets.”
GAN Augmentation: Augmenting Training Data using Generative Adversarial Networks (2018, methodology): “…formulation has been built on to address problems such as training stability [16], low resolution [10], and the absence of a true image quality based loss function [2], and applied to tasks such as super resolution [13], reconstructing images from a minimal data [22] and anomaly detection [19].”
Failing Loudly: An Empirical Study of Methods for Detecting Dataset Shift (2018, background): “Recently, also GANs have been explored for this task [39].”
Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networks (2017, background): “Schlegl et al. (2017) train a generative adversarial networks to detect out-of-distribution examples in clinical scenario.”
Synthetic data augmentation using GAN for improved liver lesion classification (2018, background): “Recently, several medical imaging applications have applied the GAN framework [6, 7, 8].”
Does Your Model Know the Digit 6 Is Not a Cat? A Less Biased Evaluation of "Outlier" Detectors (2018, background): “These methods provide an abstract idea which depends on the successful training of GANs.”
Generative Adversarial Network in Medical Imaging: A Review (2018, methodology): “On the other hand, adversarial losses as introduced by the discriminator can take into account high order potentials (Yang et al., 2017a).”
Adversarially Learned Anomaly Detection (2018, methodology): “AnoGAN [10] is the only published GAN-based method for anomaly detection.”
Classification-Based Anomaly Detection for General Data (2020, methodology): “Another set of recent methods (Schlegl et al., 2017; Deecke et al., 2018) use GANs to learn a reconstruction basis function.”
Mixed supervision for surface-defect detection: from weakly to fully supervised learning (2021, positive): “Various methods based on this principle were proposed, such as AnoGAN [31] and its successor f-AnoGAN [30] that utilize Generative Adversarial Networks, or a deep-metric-learning-based approach with triplet loss that learns features of non-anomalous samples [34], or approach that transfers pre-train”
On Data Augmentation for GAN Training (2020, background): “GAN has achieved remarkable results in various tasks, for example: image synthesis [2]–[5], image transformation [6]–[8], super-resolution [9], [10], text to image [11], [12], video captioning [13], image dehazing [14], domain adaptation [15], anomaly detection [16], [17].”
AnomalyNet: An Anomaly Detection Network for Video Surveillance (2019, background): “More recently, more and more researches focus on either of the fully convolutional neural networks (FCNs) [35] and generative adversarial networks (GANs) [36]–[39].”
A Domain Agnostic Measure for Monitoring and Evaluating GANs (2018, background): “cs.LG] 13 Nov 2018 speciﬁc to datasets of natural images, but we are also starting to notice the success of GANs in other ﬁelds such as in cosmology (Rodriguez et al., 2018) or in the medical domain (Schlegl et al., 2017). Another problem related to the evaluation of samples produced by a GAN is the”
CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instances (2020, methodology): “We compare our method with various prior methods including one-class classiﬁer [56, 53], reconstruction-based [55, 48], and self-supervised [15, 26, 1] approaches.”
Skip-GANomaly: Skip Connected and Adversarially Trained Encoder-Decoder Anomaly Detection (2019, methodology): “to perform inference within the latent space is one of the novel contributions of the proposed approach compared to the previous approaches [9], [12], [13].”