Left Ventricle Segmentation and Quantification from Cardiac Cine MR Images via Multi-task Learning

Abstract

Segmentation of the left ventricle and quantification of various cardiac contractile functions is crucial for the timely diagnosis and treatment of cardiovascular diseases. Traditionally, the two tasks have been tackled independently. Here we propose a convolutional neural network based multi-task learning approach to perform both tasks simultaneously, such that, the network learns better representation of the data with improved generalization performance. Probabilistic formulation of the problem enables learning the task uncertainties during the training, which are used to automatically compute the weights for the tasks. We performed a five fold cross-validation of the myocardium segmentation obtained from the proposed multi-task network on 97 patient 4-dimensional cardiac cine-MRI datasets available through the STA-COM LV segmentation challenge against the provided gold-standard myocardium segmentation, obtaining a Dice overlap of 0.849 ± 0.036 and mean surface distance of 0.274±0.083 mm, while simultaneously estimating the myocardial area with mean absolute difference error of 205 ± 198 mm 2 . arXiv:1809.10221v1 [cs.CV]

Deep Learning for Cardiac Image Segmentation: A Review

“…Multi-task learning: Multi-task learning has also been explored to regularize FCN-based cardiac ventricle segmentation during training by performing auxiliary tasks that are relevant to the main segmentation task, such as motion estimation (Qin et al, 2018b), estimation of cardiac function (Dangi et al, 2018b), ventricle size classification (Zhang et al, 2018b) and image reconstruction (Chartsias et al, 2018;Huang et al, 2019). Training a network for multiple tasks simultaneously encourages the network to extract features which are useful across these tasks, resulting in improved learning efficiency and prediction accuracy.…”

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

“…For segmenting the LV and myocardium from CMR images, CNNs in various orders have been proposed. Dangi et al [ 23 ] created a CNN-based multi-task learning (MTL) model for simultaneous LV segmentation and quantification. They used the U-net architecture [ 24 ], separating segmentation and regression at the final upsampling layer.…”

“…Deep learning methods have lately obtained excellent results in the segmentation of medical images. CNN is one of the most widely used methods in medical image analysis [ 23 , 43 ] among these approaches. Medical images are segmented at the pixel level, as opposed to image-level classification [ 27 ].…”

Ensemble machine learning approach for screening of coronary heart disease based on echocardiography and risk factors

“…Recently, most of the applications of the machine learning methods on echocardiogram focus on image segmentation and interpretation [16,35,36]. The methods can learn the shape and size of the region of interest from a labeled training set [39][40][41][42][43][44][45][46]. For example, machine learning methods are applied to analyzing the cardiac structures, such as determining global features that can be used to identify standard views of echocardiograms [15], extracting hidden features to detect heart diseases such as hypertrophic cardiomyopathy [16], identifying certain local structures like pacemaker lead [36], and recognizing the boundaries of ventricle and atrium [35,36].…”