Left ventricle automatic segmentation in cardiac MRI using a combined CNN and U-net approach

Wu, Bofeng; Fang, Ying; Lai, Xiaobo

doi:10.1016/j.compmedimag.2020.101719

Cited by 51 publications

(28 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moradi et al [ 25 ] developed a deep-learning-based method called MFP-U-net for LV segmentation from echocardiography images, and they designed a network with a feature pyramid that can detect and recognize the LV in MRI. Wu et al [ 26 ] proposed an automatic segmentation model for the LV from cardiac MRI. They used a CNN model to locate the LV and the U-net model to segment it.…”

Section: Related Workmentioning

confidence: 99%

Automatic Left Ventricle Segmentation from Short-Axis Cardiac MRI Images Based on Fully Convolutional Neural Network

et al. 2022

View full text Add to dashboard Cite

Background: Left ventricle (LV) segmentation using a cardiac magnetic resonance imaging (MRI) dataset is critical for evaluating global and regional cardiac functions and diagnosing cardiovascular diseases. LV clinical metrics such as LV volume, LV mass and ejection fraction (EF) are frequently extracted based on the LV segmentation from short-axis MRI images. Manual segmentation to assess such functions is tedious and time-consuming for medical experts to diagnose cardiac pathologies. Therefore, a fully automated LV segmentation technique is required to assist medical experts in working more efficiently. Method: This paper proposes a fully convolutional network (FCN) architecture for automatic LV segmentation from short-axis MRI images. Several experiments were conducted in the training phase to compare the performance of the network and the U-Net model with various hyper-parameters, including optimization algorithms, epochs, learning rate, and mini-batch size. In addition, a class weighting method was introduced to avoid having a high imbalance of pixels in the classes of image’s labels since the number of background pixels was significantly higher than the number of LV and myocardium pixels. Furthermore, effective image conversion with pixel normalization was applied to obtain exact features representing target organs (LV and myocardium). The segmentation models were trained and tested on a public dataset, namely the evaluation of myocardial infarction from the delayed-enhancement cardiac MRI (EMIDEC) dataset. Results: The dice metric, Jaccard index, sensitivity, and specificity were used to evaluate the network’s performance, with values of 0.93, 0.87, 0.98, and 0.94, respectively. Based on the experimental results, the proposed network outperforms the standard U-Net model and is an advanced fully automated method in terms of segmentation performance. Conclusion: This proposed method is applicable in clinical practice for doctors to diagnose cardiac diseases from short-axis MRI images.

show abstract

Section: Related Workmentioning

confidence: 99%

Automatic Left Ventricle Segmentation from Short-Axis Cardiac MRI Images Based on Fully Convolutional Neural Network

et al. 2022

View full text Add to dashboard Cite

show abstract

“…They reported similar segmentation performance, < 0.6 seconds per case, however the manual segmentation required 238 seconds per case. Wu et al developed a combinatorial network for segmenting the left ventricle (LV) [24]. They claimed that combining networks can increase the accuracy of the segmentation.…”

Section: Pre-processing and Correctionmentioning

confidence: 99%

Four-Dimensional Flow Magnetic Resonance Imaging and Applications in Cardiology

Geeraert¹,

Kim²,

Ali³

et al. 2022

Blood - Updates on Hemodynamics and Thalassemia

View full text Add to dashboard Cite

Blood flow through the heart and great vessels moves in three dimensions (3D) throughout time. However, the assessment of its 3D nature has been limited in the human body. Recent advances in magnetic resonance imaging (MRI) allow for the comprehensive visualization and quantification of in-vivo flow dynamics using four-dimensional (4D) flow MRI. In addition, this technique provides the opportunity to obtain advanced hemodynamic biomarkers such as vorticity, helicity, wall shear stress (WSS), pressure gradients, viscous energy loss (EL), and turbulent kinetic energy (TKE). This chapter will introduce 4D flow MRI which is currently used for blood flow visualization and advanced quantification of cardiac hemodynamic biomarkers. We will discuss its advantages relative to other in-vivo flow imaging techniques and describe its potential clinical applications in cardiology.

show abstract

“…Combined CNN (Pereira et al, 2020) and U-Net method (B. Wu, Fang, & Lai, 2020;Wu et al, 2021) detects borders of left ventricle in magnetic resonance images. Due to limitation of CNN, adversarial learning (Chen et al, 2019;Wu, Lu, et al, 2021) segmentation method is tested.…”

Section: Related Workmentioning

confidence: 99%

Discrete light sheet microscopic segmentation of left ventricle using morphological tuning and active contours

Irshad

Sharif

Yasmin

et al. 2021

Microscopy Res & Technique

View full text Add to dashboard Cite

show abstract

Left ventricle automatic segmentation in cardiac MRI using a combined CNN and U-net approach

Cited by 51 publications

References 17 publications

Automatic Left Ventricle Segmentation from Short-Axis Cardiac MRI Images Based on Fully Convolutional Neural Network

Automatic Left Ventricle Segmentation from Short-Axis Cardiac MRI Images Based on Fully Convolutional Neural Network

Four-Dimensional Flow Magnetic Resonance Imaging and Applications in Cardiology

Discrete light sheet microscopic segmentation of left ventricle using morphological tuning and active contours

Contact Info

Product

Resources

About