Deep Autoencoding Models for Unsupervised Anomaly Segmentation in Brain MR Images

Abstract

Reliably modeling normality and differentiating abnormal appearances from normal cases is a very appealing approach for detecting pathologies in medical images. A plethora of such unsupervised anomaly detection approaches has been made in the medical domain, based on statistical methods, content-based retrieval, clustering and recently also deep learning. Previous approaches towards deep unsupervised anomaly detection model patches of normal anatomy with variants of Autoencoders or GANs, and detect anomalies either as outliers in the learned feature space or from large reconstruction errors. In contrast to these patch-based approaches, we show that deep spatial autoencoding models can be efficiently used to capture normal anatomical variability of entire 2D brain MR images. A variety of experiments on real MR data containing MS lesions corroborates our hypothesis that we can detect and even delineate anomalies in brain MR images by simply comparing input images to their reconstruction. Results show that constraints on the latent space and adversarial training can further improve the segmentation performance over standard deep representation learning.

References in corpus (4)

• Auto-Encoding Variational Bayes
• Adversarial Feature Learning
• Variational Approaches for Auto-Encoding Generative Adversarial Networks
• Identifying and Categorizing Anomalies in Retinal Imaging Data

Cited by in corpus (70)

• Uninformed Students: Student-Teacher Anomaly Detection with Discriminative Latent Embeddings
• Improving Unsupervised Defect Segmentation by Applying Structural Similarity to Autoencoders
• Deep Learning for Unsupervised Anomaly Localization in Industrial Images: A Survey
• DFR: Deep Feature Reconstruction for Unsupervised Anomaly Segmentation
• Scaling Out-of-Distribution Detection for Real-World Settings
• RA-GCN: Graph Convolutional Network for Disease Prediction Problems with Imbalanced Data
• Student-Teacher Feature Pyramid Matching for Anomaly Detection
• Anomaly Detection in Medical Imaging -- A Mini Review
• Deep and Statistical Learning in Biomedical Imaging: State of the Art in 3D MRI Brain Tumor Segmentation
• Context-encoding Variational Autoencoder for Unsupervised Anomaly Detection
• Machine Learning in NextG Networks via Generative Adversarial Networks
• Pseudo-healthy synthesis with pathology disentanglement and adversarial learning
• Attention-based Convolutional Autoencoders for 3D-Variational Data Assimilation
• Unsupervised brain lesion segmentation from MRI using a convolutional autoencoder
• Anomaly Detection based on Zero-Shot Outlier Synthesis and Hierarchical Feature Distillation
• AnoSeg: Anomaly Segmentation Network Using Self-Supervised Learning
• Deep Generative Models in the Real-World: An Open Challenge from Medical Imaging
• Masked Transformer for image Anomaly Localization
• Unsupervised Brain Anomaly Detection and Segmentation with Transformers
• Unsupervised Pathology Detection: A Deep Dive Into the State of the Art
• Towards reliable head and neck cancers locoregional recurrence prediction using delta-radiomics and learning with rejection option
• Challenging Current Semi-Supervised Anomaly Segmentation Methods for Brain MRI
• PyOD: A Python Toolbox for Scalable Outlier Detection
• Encoding Structure-Texture Relation with P-Net for Anomaly Detection in Retinal Images
• A Case for the Score: Identifying Image Anomalies using Variational Autoencoder Gradients
• Noise Reduction in X-ray Photon Correlation Spectroscopy with Convolutional Neural Networks Encoder-Decoder Models
• Unsupervised 3D Brain Anomaly Detection
• Reconstruction Student with Attention for Student-Teacher Pyramid Matching
• ASC-Net : Adversarial-based Selective Network for Unsupervised Anomaly Segmentation
• Simultaneous imputation and disease classification in incomplete medical datasets using Multigraph Geometric Matrix Completion (MGMC)
• Robust Variational Autoencoder
• Puzzle-AE: Novelty Detection in Images through Solving Puzzles
• Robust Image Segmentation Quality Assessment
• Unsupervised anomaly localization using VAE and beta-VAE
• Weakly Supervised Attention Model for RV StrainClassification from volumetric CTPA Scans
• Variational AutoEncoder For Regression: Application to Brain Aging Analysis
• Synthesize then Compare: Detecting Failures and Anomalies for Semantic Segmentation
• ICAM: Interpretable Classification via Disentangled Representations and Feature Attribution Mapping
• One Versus all for deep Neural Network Incertitude (OVNNI) quantification
• Brain Tumor Anomaly Detection via Latent Regularized Adversarial Network
• Evaluation of pseudo-healthy image reconstruction for anomaly detection with deep generative models: Application to brain FDG PET
• Unsupervised Anomaly Segmentation using Image-Semantic Cycle Translation
• Self-Supervised Out-of-Distribution Detection in Brain CT Scans
• Image Synthesis as a Pretext for Unsupervised Histopathological Diagnosis
• High- and Low-level image component decomposition using VAEs for improved reconstruction and anomaly detection
• Comparison of Outlier Detection Techniques for Structured Data
• Pixel-wise Energy-biased Abstention Learning for Anomaly Segmentation on Complex Urban Driving Scenes
• Label-Free Segmentation of COVID-19 Lesions in Lung CT
• Addressing Variance Shrinkage in Variational Autoencoders using Quantile Regression
• Generative Models for Reproducible Coronary Calcium Scoring
• Region and Spatial Aware Anomaly Detection for Fundus Images
• Variational Autoencoders with a Structural Similarity Loss in Time of Flight MRAs
• Walking the Tightrope: An Investigation of the Convolutional Autoencoder Bottleneck
• Medical Knowledge-Guided Deep Learning for Imbalanced Medical Image Classification
• PNUNet: Anomaly Detection using Positive-and-Negative Noise based on Self-Training Procedure
• Deep Metric Learning for Open World Semantic Segmentation
• Attention Map-guided Two-stage Anomaly Detection using Hard Augmentation
• Convolutional Recurrent Reconstructive Network for Spatiotemporal Anomaly Detection in Solder Paste Inspection
• Quantum Patch-Based Autoencoder for Anomaly Segmentation
• Progressive GANomaly: Anomaly detection with progressively growing GANs
• Teacher Encoder-Student Decoder Denoising Guided Segmentation Network for Anomaly Detection
• Unsupervised Anomaly Detection in MR Images using Multi-Contrast Information
• Spoof Face Detection Via Semi-Supervised Adversarial Training
• Learning Non-Unique Segmentation with Reward-Penalty Dice Loss
• Unsupervised View-Invariant Human Posture Representation
• 3-Dimensional Deep Learning with Spatial Erasing for Unsupervised Anomaly Segmentation in Brain MRI
• 3D Reasoning for Unsupervised Anomaly Detection in Pediatric WbMRI
• Open-set 3D Object Detection
• Semi-Supervised Surface Anomaly Detection of Composite Wind Turbine Blades From Drone Imagery
• Improved Slice-wise Tumour Detection in Brain MRIs by Computing Dissimilarities between Latent Representations