Deep Learning for Segmentation Using an Open Large-Scale Dataset in 2D Echocardiography

Abstract

Delineation of the cardiac structures from 2D echocardiographic images is a common clinical task to establish a diagnosis. Over the past decades, the automation of this task has been the subject of intense research. In this paper, we evaluate how far the state-of-the-art encoder-decoder deep convolutional neural network methods can go at assessing 2D echocardiographic images, i.e. segmenting cardiac structures as well as estimating clinical indices, on a dataset especially designed to answer this objective. We therefore introduce the Cardiac Acquisitions for Multi-structure Ultrasound Segmentation (CA-MUS) dataset, the largest publicly-available and fully-annotated dataset for the purpose of echocardiographic assessment. The dataset contains two and four-chamber acquisitions from 500 patients with reference measurements from one cardiologist on the full dataset and from three cardiologists on a fold of 50 patients. Results show that encoder-decoder based architectures outperform state-of-the-art non-deep learning methods and faithfully reproduce the expert analysis for the end-diastolic and endsystolic left ventricular volumes, with a mean correlation of 0.95 and an absolute mean error of 9.5 ml. Concerning the ejection fraction of the left ventricle, results are more contrasted with a mean correlation coefficient of 0.80 and an absolute mean error of 5.6%. Although these results are below the inter-observer scores, they remain slightly worse than the intra-observer's ones. Based on this observation, areas for improvement are defined, which open the door for accurate and fully-automatic analysis of 2D echocardiographic images.Index Terms-Cardiac segmentation and diagnosis, deep learning, ultrasound, left ventricle, myocardium, left atrium.

Deep Learning for Cardiac Image Segmentation: A Review

“…The dataset is composed of apical 4-chamber view images annotated for LV and LA segmentation, with uneven imaging quality from 500 patients with varying conditions. Notably, the initial benchmarking (Leclerc et al, 2019) on this dataset has shown that modern encoder-decoder CNNs resulted in lower error than inter-observer error between human cardiologists. Ghesu et al (2016) proposed a framework based on marginal space learning (MSL), deep neural networks (DNNs) and active shape model (ASM) to segment the aortic valve in 3D cardiac ultrasound volumes.…”

Recurrent Aggregation Learning for Multi-view Echocardiographic Sequences Segmentation

“…To validate the efficiency of RAL, we built a large multi-view echocardiographic sequences dataset, which was acquired from three centers' various vendor machines (The Second People's Hospital of Shenzhen, The Third People's Hospital of Shenzhen, and Peking University First Hospital). We further evaluate RAL on the public CAMUS dataset [6]. Our dataset contains 300 sequences from 3 views and every sequence includes 30 frames.…”

“…The application scenario of existing methods is always limited and only suitable under a specific situation. They mostly focus on specific view [4] or single frames (i.e., without considering the sequence) [5] or one single vendor and center [6]. As for sequence segmentation, existing methods try to leverage temporal information by using a deformable model combined with the optical flow [7,8] or fine-tuning pretrained CNN dynamically with first frame's label till the last frame [9].…”

Segmentation of Left Ventricle in 2D Echocardiography Using Deep Learning

“…The number of feature maps given in table corresponds to the number of convolutions in the convolution layers. For each U-Net implementation, the values for the first, the bottom (where the spatial information is the most compressed), and the last convolution layers indicated (Leclerc et al, 2019).…”

“…There are many suggested methods for 2D LV segmentation. Recently Deep Convolutional Neural Networks have shown promising results for image segmentation (Jafari et al, 2018;Leclerc et al, 2019), specifically U-Net, which has been successfully applied to multiple medical image segmentation problems. As explained in the abstract, this study aims to investigate and compare the performance of U-Net 1 and U-Net 2 models reported by (Leclerc et al, 2019) with the original U-Net (Ronneberger et al, 2015).…”