• Corpus ID: 237571569

Deep Quantile Regression for Uncertainty Estimation in Unsupervised and Supervised Lesion Detection

  title={Deep Quantile Regression for Uncertainty Estimation in Unsupervised and Supervised Lesion Detection},
  author={Haleh Akrami and Anand A. Joshi and Serg{\"u}l Ayd{\"o}re and Richard Leahy},
Despite impressive state-of-the-art performance on a wide variety of machine learning tasks in multiple applications, deep learning methods can produce over-confident predictions, particularly with limited training data. Therefore, quantifying uncertainty is particularly important in critical applications such as lesion detection and clinical diagnosis, where a realistic assessment of uncertainty is essential in determining surgical margins, disease status and appropriate treatment. In this work… 

Figures and Tables from this paper



Unsupervised Lesion Detection via Image Restoration with a Normative Prior

A probabilistic model that uses a network-based prior as the normative distribution and detect lesions pixel-wise using MAP estimation is proposed and outperforms the state-of-the-art unsupervised methods by a substantial margin.

Brain Lesion Detection Using A Robust Variational Autoencoder and Transfer Learning

The results on MRI datasets demonstrate that the method proposed can improve the accuracy of lesion detection by adapting robust statistical models and transfer learning for a variational autoencoder model.

Leveraging Uncertainty Estimates for Predicting Segmentation Quality

This paper proposes a two-stage architecture for producing measures of uncertainty, which can accommodate any deep learning-based medical segmentation pipeline and shows that reasoning about spatial uncertainty, the first output, is a useful intermediate representation for generating segmentation quality predictions, the second output.

Unsupervised Detection of Lesions in Brain MRI using constrained adversarial auto-encoders

This work uses the Human Connectome Project dataset to learn distribution of healthy-appearing brain MRI and proposes a simple yet effective constraint that helps mapping of an image bearing lesion close to its corresponding healthy image in the latent space.

Deep Autoencoding Models for Unsupervised Anomaly Segmentation in Brain MR Images

This work shows that deep spatial autoencoding models can be efficiently used to capture normal anatomical variability of entire 2D brain MR images and shows that constraints on the latent space and adversarial training can further improve the segmentation performance over standard deep representation learning.

Stochastic Segmentation Networks: Modelling Spatially Correlated Aleatoric Uncertainty

Stochastic segmentation networks (SSNs) are introduced, an efficient probabilistic method for modelling aleatoric uncertainty with any image segmentation network architecture and outperform state-of-the-art for modelling correlated uncertainty in ambiguous images while being much simpler, more flexible, and more efficient.

Unsupervised Lesion Detection in Brain CT using Bayesian Convolutional Autoencoders

This work explores the use of Bayesian autoencoders to learn the variability of healthy tissue and detect lesions as unlikely events under the normative model and results indicate that the method achieves best performance in detecting lesions caused by bleeding compared to baselines.

Supervised Uncertainty Quantification for Segmentation with Multiple Annotations

This work exploits multi-grader annotation variability as a source of 'groundtruth' aleatoric uncertainty, which can be treated as a target in a supervised learning problem and combined with a Probabilistic U-Net is able to improve predictive uncertainty estimates.

A Probabilistic U-Net for Segmentation of Ambiguous Images

A generative segmentation model based on a combination of a U-Net with a conditional variational autoencoder that is capable of efficiently producing an unlimited number of plausible hypotheses and reproduces the possible segmentation variants as well as the frequencies with which they occur significantly better than published approaches.

Spatially Varying Label Smoothing: Capturing Uncertainty from Expert Annotations

Spatially Varying Label Smoothing (SVLS) is proposed, a soft labeling technique that captures the structural uncertainty in semantic segmentation and naturally lends itself to incorporate inter-rater uncertainty when multiple labelmaps are available.