Hierarchical LoG Bayesian Neural Network for Enhanced Aorta Segmentation

TL;DR

This work addresses the challenge of segmenting the aorta and its supra-aortic branches from multiscale CT angiography images while quantify­ing prediction uncertainty. It introduces LoGB-Net, a dual-stream architecture that combines a regular 3D U-Net with a hierarchical Laplacian of Gaussian LoG stream. The LoG branch is Bayesianly parameterized, with trainable LoG kernels across multiple scales and a balanced gate to address foreground-background imbalance, enabling self-adaptive vessel enhancement and uncertainty quantification via the ELBO objective $\ abla$ $p(\theta|D)$. Empirically, LoGB-Net outperforms baselines on two CTA datasets, achieving about a 3% Dice gain, and provides confidence intervals for segmentation that inform downstream tasks such as CFD simulations. The approach offers robust aorta segmentation with practical uncertainty estimates, supporting more reliable vascular analysis in clinical research and planning.

Abstract

Accurate segmentation of the aorta and its associated arch branches is crucial for diagnosing aortic diseases. While deep learning techniques have significantly improved aorta segmentation, they remain challenging due to the intricate multiscale structure and the complexity of the surrounding tissues. This paper presents a novel approach for enhancing aorta segmentation using a Bayesian neural network-based hierarchical Laplacian of Gaussian (LoG) model. Our model consists of a 3D U-Net stream and a hierarchical LoG stream: the former provides an initial aorta segmentation, and the latter enhances blood vessel detection across varying scales by learning suitable LoG kernels, enabling self-adaptive handling of different parts of the aorta vessels with significant scale differences. We employ a Bayesian method to parameterize the LoG stream and provide confidence intervals for the segmentation results, ensuring robustness and reliability of the prediction for vascular medical image analysts. Experimental results show that our model can accurately segment main and supra-aortic vessels, yielding at least a 3% gain in the Dice coefficient over state-of-the-art methods across multiple volumes drawn from two aorta datasets, and can provide reliable confidence intervals for different parts of the aorta. The code is available at https://github.com/adlsn/LoGBNet.

Figures (5)

• Figure 1: The top shows the flow diagram of LoGB-Net. The Bayesian segmentation results' blue, red, and green lines represent the segmentation's upper bound, mean, and lower bound. The bottom displays the application of UQ in aorta segmentation for downstream tasks, such as CFD simulation.
• Figure 2: Confidence intervals for aortic predictions based on boundary clarity variations.
• Figure 3: Qualitative results of different methods for one volume in the testing set (drawn from the first dataset). The top row displays the 3D segmentation prediction, where selected slices (S1 to S2) are marked on the ground-truth result.
• Figure 4: CFD simulation results for the aortic vessel's velocity, pressure, and wall shear stress. Each case represents a different UQ realization.
• Figure 5: Qualitative results of the LoGB-Net's ablation study.