DeepVesselNet: Vessel Segmentation, Centerline Prediction, and Bifurcation Detection in 3-D Angiographic Volumes

03/25/2018
by   Giles Tetteh, et al.
0

We present DeepVesselNet, an architecture tailored to the challenges to be addressed when extracting vessel networks and corresponding features in 3-D angiography using deep learning. We discuss the problems of low execution speed and high memory requirements associated with full 3-D convolutional networks, high class imbalance arising from low percentage (less than 3 voxels, and unavailability of accurately annotated training data - and offer solutions that are the building blocks of DeepVesselNet. First, we formulate 2-D orthogonal cross-hair filters which make use of 3-D context information. Second, we introduce a class balancing cross-entropy score with false positive rate correction to handle the high class imbalance and high false positive rate problems associated with existing loss functions. Finally, we generate synthetic dataset using a computational angiogenesis model, capable of generating vascular networks under physiological constraints on local network structure and topology, and use these data for transfer learning. DeepVesselNet is optimized for segmenting vessels, predicting centerlines, and localizing bifurcations. We test the performance on a range of angiographic volumes including clinical Time-of-Flight MRA data of the human brain, as well as synchrotron radiation X-ray tomographic microscopy scans of the rat brain. Our experiments show that, by replacing 3-D filters with 2-D orthogonal cross-hair filters in our network, speed is improved by 23 maintained. Our class balancing metric is crucial for training the network and pre-training with synthetic data helps in early convergence of the training process.

READ FULL TEXT

page 20

page 29

page 30

research
02/21/2021

Constrained Optimization for Training Deep Neural Networks Under Class Imbalance

Deep neural networks (DNNs) are notorious for making more mistakes for t...
research
02/01/2021

Segmenting Microcalcifications in Mammograms and its Applications

Microcalcifications are small deposits of calcium that appear in mammogr...
research
08/07/2023

Unsupervised Adversarial Detection without Extra Model: Training Loss Should Change

Adversarial robustness poses a critical challenge in the deployment of d...
research
04/06/2019

DeepSEED: 3D Squeeze-and-Excitation Encoder-Decoder ConvNets for Pulmonary Nodule Detection

Pulmonary nodule detection plays an important role in lung cancer screen...
research
12/22/2021

Deep learning for brain metastasis detection and segmentation in longitudinal MRI data

Brain metastases occur frequently in patients with metastatic cancer. Ea...
research
11/18/2018

RePr: Improved Training of Convolutional Filters

A well-trained Convolutional Neural Network can easily be pruned without...

Please sign up or login with your details

Forgot password? Click here to reset