Learning Wall Segmentation in 3D Vessel Trees using Sparse Annotations
This addresses the challenge of creating 3D medical image segmentations with limited annotations for clinicians, though it is incremental as it builds on existing adversarial and 3D CNN techniques.
The paper tackles the problem of 3D carotid artery wall segmentation by using sparse 2D annotations to generate 3D pseudo-labels, circumventing the need for manual 3D masks. The result is an efficient training method that enhances segmentation performance, particularly in bifurcation areas, with potential applications in stenosis assessment and 3D biomarker extraction.
We propose a novel approach that uses sparse annotations from clinical studies to train a 3D segmentation of the carotid artery wall. We use a centerline annotation to sample perpendicular cross-sections of the carotid artery and use an adversarial 2D network to segment them. These annotations are then transformed into 3D pseudo-labels for training of a 3D convolutional neural network, circumventing the creation of manual 3D masks. For pseudo-label creation in the bifurcation area we propose the use of cross-sections perpendicular to the bifurcation axis and show that this enhances segmentation performance. Different sampling distances had a lesser impact. The proposed method allows for efficient training of 3D segmentation, offering potential improvements in the assessment of carotid artery stenosis and allowing the extraction of 3D biomarkers such as plaque volume.