Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Bayesian Neural Networks for 2D MRI Segmentation

Lohith Konathala

TL;DR

BA U-Net is introduced, an uncertainty-aware model for MRI segmentation that integrates Bayesian Neural Networks with Attention Mechanisms that addresses the critical need for confidence estimation in deep learning-based medical imaging.

Abstract

Uncertainty quantification is vital for safety-critical Deep Learning applications like medical image segmentation. We introduce BA U-Net, an uncertainty-aware model for MRI segmentation that integrates Bayesian Neural Networks with Attention Mechanisms. BA U-Net delivers accurate, interpretable results, crucial for reliable pathology screening. Evaluated on BraTS 2020, this model addresses the critical need for confidence estimation in deep learning-based medical imaging.

Bayesian Neural Networks for 2D MRI Segmentation

TL;DR

BA U-Net is introduced, an uncertainty-aware model for MRI segmentation that integrates Bayesian Neural Networks with Attention Mechanisms that addresses the critical need for confidence estimation in deep learning-based medical imaging.

Abstract

Uncertainty quantification is vital for safety-critical Deep Learning applications like medical image segmentation. We introduce BA U-Net, an uncertainty-aware model for MRI segmentation that integrates Bayesian Neural Networks with Attention Mechanisms. BA U-Net delivers accurate, interpretable results, crucial for reliable pathology screening. Evaluated on BraTS 2020, this model addresses the critical need for confidence estimation in deep learning-based medical imaging.
Paper Structure (16 sections, 9 equations, 7 figures, 2 tables)

This paper contains 16 sections, 9 equations, 7 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Image Segmentation
  4. Uncertainty Quantification
  5. Methodology
  6. Model Architecture
  7. Design Choices
  8. Experimental Setup
  9. Dataset
  10. Model Training
  11. Model Inference
  12. Results
  13. Gaussian Blur
  14. Rician Noise
  15. Discussion
  16. ...and 1 more sections

Figures (7)

  • Figure 1: Bayesian Attention U-Net Decoder Block
  • Figure 2: Architecture of Bayesian Attention U-Net
  • Figure 3: BraTS 2020 T2-FLAIR MRI section with ground-truth tumour annotation
  • Figure 4: Variance Analysis of Aleatoric and Epistemic Uncertainties
  • Figure 5: BA U-Net Inference on BraTS 2020 Dataset
  • ...and 2 more figures