Abstract:Background: Late gadolinium enhancement cardiac magnetic resonance imaging (LGE-CMR) is a valuable cardiovascular imaging technique. Segmentation of cardiac chambers from LGE-CMR is a fundamental step in electrophysiological modeling and cardiovascular disease diagnosis. Deep learning methods have demonstrated extremely promising performance. However, excellent performance often depended on a large amount of finely annotated data. The purpose of this manuscript was to develop a semi-supervised segmentation method to use unlabeled data to improve model performance. Methods: This manuscript proposed a semi-supervised network that integrates triple-consistency constraints (data-level, task-level, and feature-level) for cardiac chambers segmentation from LGE-CMR. Specifically, we designed a network that integrated segmentation and edge prediction tasks based on the mean teacher architecture. This addressed the problem of ignoring some challenging regions because of excluding low-confidence regions of previous research. We also applied a voxel-level contrastive learning strategy to achieve feature-level consistency, helping the model pay attention to the consistency between features overlooked in previous research. Results: In terms of the Dice, Jaccard, Average Surface Distance (ASD), and 95% Hausdorff Distance (95HD) metrics, for the atrial segmentation dataset, the proposed method achieved scores of 88.34%, 79.30%, 7.92, and 2.02 when trained with 10% labeled data, and 90.70%, 83.09%, 6.41, and 1.72 when trained with 20% labeled data. For the ventricular segmentation task, the results were 87.22%, 77.95%, 2.27, and 0.61 with 10% labeled data, and 88.99%, 80.45%, 1.87, and 0.51 with 20% labeled data, respectively. Conclusion: Experiments demonstrated that our method outperforms previous semi-supervised methods, showing the potential of the proposed network for semi-supervised segmentation problems.

Hierarchical Consistency Regularized Mean Teacher for Semi-supervised 3D Left Atrium Segmentation

Hierarchical Consistency and Refinement for Semi-supervised Medical Segmentation.

Segmentation of Cardiac Tissues and Organs for CCTA Images Based on a Deep Learning Model

Competitive Ensembling Teacher-Student Framework for Semi-Supervised Left Atrium MRI Segmentation

DC-net: Dual-Consistency semi-supervised learning for 3D left atrium segmentation from MRI

3D Graph-S<SUP>2</SUP>Net: Shape-Aware Self-ensembling Network for Semi-supervised Segmentation with Bilateral Graph Convolution

MLC: Multi-level consistency learning for semi-supervised left atrium segmentation

Semi-supervised Left Atrium Segmentation with Mutual Consistency Training

Semi-supervised segmentation of cardiac chambers from LGE-CMR using feature consistency awareness

Complementary consistency semi-supervised learning for 3D left atrial image segmentation

Automatic Delineation of the 3D Left Atrium from LGE-MRI: Actor-Critic based Detection and Semi-Supervised Segmentation

DTSC-Net: Semi-supervised 3D Biomedical Image Segmentation through Dual-Teacher Simplified Consistency.

Semi-supervised Histological Image Segmentation Via Hierarchical Consistency Enforcement

Synergistic Label-Stability Learning for Semi-supervised Left Atrium Segmentation

Constantly optimized mean teacher for semi-supervised 3D MRI image segmentation

Exploring Inherent Consistency for Semi-supervised Anatomical Structure Segmentation in Medical Imaging

Semi-supervised Medical Image Segmentation via Geometry-aware Consistency Training

A Teacher-Student Framework for Semi-supervised Medical Image Segmentation From Mixed Supervision

Segmentation and visualization of left atrium through a unified deep learning framework

Triple-task mutual consistency for semi-supervised 3D medical image segmentation

3D Vessel Segmentation with Limited Guidance of 2D Structure-agnostic Vessel Annotations