A Case for the Score: Identifying Image Anomalies using Variational Autoencoder Gradients
arXiv:1912.00003
Abstract
Through training on unlabeled data, anomaly detection has the potential to impact computer-aided diagnosis by outlining suspicious regions. Previous work on deep-learning-based anomaly detection has primarily focused on the reconstruction error. We argue instead, that pixel-wise anomaly ratings derived from a Variational Autoencoder based score approximation yield a theoretically better grounded and more faithful estimate. In our experiments, Variational Autoencoder gradient-based rating outperforms other approaches on unsupervised pixel-wise tumor detection on the BraTS-2017 dataset with a ROC-AUC of 0.94.
References in corpus (6)
- Auto-Encoding Variational Bayes
- SmoothGrad: removing noise by adding noise
- Detecting Cancer Metastases on Gigapixel Pathology Images
- Deep Autoencoding Models for Unsupervised Anomaly Segmentation in Brain MR Images
- Glow: Generative Flow with Invertible 1x1 Convolutions
- Deep Generative Models in the Real-World: An Open Challenge from Medical Imaging
Cited by in corpus (7)
- Deep Learning for Unsupervised Anomaly Localization in Industrial Images: A Survey
- Iterative energy-based projection on a normal data manifold for anomaly localization
- Unsupervised anomaly localization using VAE and beta-VAE
- Brain Tumor Anomaly Detection via Latent Regularized Adversarial Network
- Stack of discriminative autoencoders for multiclass anomaly detection in endoscopy images
- Improved Slice-wise Tumour Detection in Brain MRIs by Computing Dissimilarities between Latent Representations
- Context-encoding Variational Autoencoder for Unsupervised Anomaly Detection -- Short Paper