Automated Segmentation on the Entire Cardiac Cycle Using a Deep Learning Work - Flow

Abstract

The segmentation of the left ventricle (LV) from CINE MRI images is essential to infer important clinical parameters. Typically, machine learning algorithms for automated LV segmentation use annotated contours from only two cardiac phases, diastole, and systole. In this work, we present an analysis work-flow for fully-automated LV segmentation that learns from images acquired through the cardiac cycle. The workflow consists of three components: first, for each image in the sequence, we perform an automated localization and subsequent cropping of the bounding box containing the cardiac silhouette. Second, we identify the LV contours using a Temporal Fully Convolutional Neural Network (T-FCNN), which extends Fully Convolutional Neural Networks (FCNN) through a recurrent mechanism enforcing temporal coherence across consecutive frames. Finally, we further defined the boundaries using either one of two components: fully-connected Conditional Random Fields (CRFs) with Gaussian edge potentials and Semantic Flow. Our initial experiments suggest that significant improvement in performance can potentially be achieved by using a recurrent neural network component that explicitly learns cardiac motion patterns whilst performing LV segmentation.

Implementation and Validation of a Three-dimensional Cardiac Motion Estimation Network

“…However, there was no significant difference between the DSCs obtained across all methods and CarMEN except for Vampire, which had a significantly lower DSC (median, 0. 33 There was no significant difference in the mean squared error and peak signal-to-noise ratio metrics. Representative images of the four patient groups for a B-spline method, dDemons, Vampire, and CarMEN are shown in Figure 3.…”

“…Although accounting for these differences was beyond the scope of our work, a novel learning-based solution for automatic intersection motion detection and correction recently proposed by Tarroni et al (32) could be incorporated as a preprocessing step. In fact, although the current implementation of CarMEN is semiautomated due to the required manual centering and cropping of the left ventricle, automatic convolutional neural network-based techniques exist to automate this process (33). However, cropping could potentially limit motion compensated reconstructions where motion estimates for the entire field of view are needed.…”

Left Ventricle Segmentation in Cardiac MR: A Systematic Mapping of the Past Decade

“…The analysis of connected components can be applied in the three-dimensional domain to ensure consistency between segmentations of diferent slices [45]. In [90], energy optimization methods based on conditional random ields and semantic low are applied to segment diferent frames, reinforcing the temporal consistency.…”

“…The initial contour is used in the initialization of a more precise method, such as a DM [119] or a more speciic AI model [103]. The sub-image is primarily obtained based on regression, in which the coordinates of the delimiting rectangle are estimated [5,38,72,90,101,126,127]. Regression is also used to obtain the center of the LV [98].…”

“…The resulting masks and original images are collected in volumes and used as input to a 3D U-net, which rectify the segmentation. [90] add a recurring layer that is able to retain information about previously segmented frames, promoting temporal coherence. In the approach presented in [128], the contextual information of slices in sequence is used in a model adapted from the U-net, which receives not only the image of a slice, but also the image and the segmentation from the previous slice.…”