Effective extraction of ventricles and myocardium objects from cardiac magnetic resonance images with a multi-task learning U-Net

•A U-Net based Multi-task deep learning framework (MTL-UNet) was proposed.•An edge detection module and fusion module are combined for improved segmentation.•Combined loss functions for accurate extraction of various semantic objects in MRI.•Improved results on extraction of ventricles and myocardiu...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	Pattern recognition letters 2022-03, Vol.155, p.165-170
Hauptverfasser:	Ren, Jinchang, Sun, He, Zhao, Huimin, Gao, Hao, Maclellan, Calum, Zhao, Sophia, Luo, Xiaoyu
Format:	Artikel
Sprache:	eng
Schlagworte:	Coronary artery disease Deep learning Entropy Fusion-based decoder Heart diseases Image segmentation Magnetic resonance images (MRI) Magnetic resonance imaging Medical imaging Modules Multi-task learning Myocardium Object recognition U-Net Ventricles and myocardium extraction
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!