Direct and Simultaneous Four-Chamber Volume Estimation by Multi-Output Regression

Zhen, Xiantong; Islam, Ali; Bhaduri, Mousumi; Chan, Ian; Li, Shuo

doi:10.1007/978-3-319-24553-9_82

Cited by 36 publications

(32 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Two-phase framework employed in existing direct methods [6]- [11] is inadequate to achieve accurate estimation for Right: the cardiac indices to be estimated in this work, which include area of LV cavity (A1), area of myocardium (A2), and the 6 regional wall thickness (WT1∼WT6). multitype cardiac indices, for the reason that image representation and indices regression are separately handled, and no feedback connection exists between them during optimization.…”

Section: B Existing Two-phase Direct Methodsmentioning

confidence: 99%

“…These handcrafted features were further replaced with a more effective image representation that learned from a multiscale convolutional deep belief network (MCDBN) [10], leading to improved correlation between the estimated volumes and their ground truth. Supervised descriptor learning (SDL) was proposed in the work of four chamber volumes estimation [11], which still employed a separate adaptive K-clustering random forest (AKRF) regression [16]. The compatibility between the descriptor and the regression model still can not be enhanced in the two-phase framework.…”

Section: B Existing Two-phase Direct Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Direct Multitype Cardiac Indices Estimation via Joint Representation and Regression Learning

Xue

Islam

Bhaduri

et al. 2017

IEEE Trans. Med. Imaging

Self Cite

View full text Add to dashboard Cite

Cardiac indices estimation is of great importance during identification and diagnosis of cardiac disease in clinical routine. However, estimation of multitype cardiac indices with consistently reliable and high accuracy is still a great challenge due to the high variability of cardiac structures and the complexity of temporal dynamics in cardiac MR sequences. While efforts have been devoted into cardiac volumes estimation through feature engineering followed by a independent regression model, these methods suffer from the vulnerable feature representation and incompatible regression model. In this paper, we propose a semi-automated method for multitype cardiac indices estimation. After the manual labeling of two landmarks for ROI cropping, an integrated deep neural network Indices-Net is designed to jointly learn the representation and regression models. It comprises two tightly-coupled networks, such as a deep convolution autoencoder for cardiac image representation, and a multiple output convolution neural network for indices regression. Joint learning of the two networks effectively enhances the expressiveness of image representation with respect to cardiac indices, and the compatibility between image representation and indices regression, thus leading to accurate and reliable estimations for all the cardiac indices. When applied with five-fold cross validation on MR images of 145 subjects, Indices-Net achieves consistently low estimation error for LV wall thicknesses (1.44 ± 0.71 mm) and areas of cavity and myocardium (204 ± 133 mm). It outperforms, with significant error reductions, segmentation method (55.1% and 17.4%), and two-phase direct volume-only methods (12.7% and 14.6%) for wall thicknesses and areas, respectively. These advantages endow the proposed method a great potential in clinical cardiac function assessment.

show abstract

Section: B Existing Two-phase Direct Methodsmentioning

confidence: 99%

Section: B Existing Two-phase Direct Methodsmentioning

confidence: 99%

Direct Multitype Cardiac Indices Estimation via Joint Representation and Regression Learning

Xue

Islam

Bhaduri

et al. 2017

IEEE Trans. Med. Imaging

Self Cite

View full text Add to dashboard Cite

show abstract

“…In these methods, hand-crafted features extracted from cardiac images are fed into regression models such as random forest (RF), adaptive K-clustering RF (AKRF), Bayesian model, and neural networks, to predict cardiac volumes. The employed features include Bhattacharyya coefficient between image distributions [1,2], appearance features [18], multiple low level image features [21], as well as features from multiscale convolutional deep belief network (MCDNB) [22] and supervised descriptor learning (SDL) [20]. Although these methods obtained effective performance, two limitations still exist: 1) they followed two separated phases, i.e., feature extraction+ volumes regression, and no feedback exists between them to make them compatible with each other; 2) neither the temporal dependencies nor the spatial dependencies are taken into account, while the dependencies are important for dynamic modeling of cardiac sequence.…”

Section: Related Workmentioning

confidence: 99%

“…To demonstrate the advantages of our proposed method over segmentation based [3] and two-phase direct methods [21,22,20], we apply these methods to our database for cardiac RWT estimation. For the direct methods, the same five-fold crossvalidation protocol is used.…”

Section: Performance Comparison: Resrnn Vs State-of-the-artmentioning

confidence: 99%

Direct Estimation of Regional Wall Thicknesses via Residual Recurrent Neural Network

Xue

Nachum

Pandey

et al. 2017

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Accurate estimation of regional wall thicknesses (RWT) of left ventricular (LV) myocardium from cardiac MR sequences is of significant importance for identification and diagnosis of cardiac disease. Existing RWT estimation still relies on segmentation of LV myocardium, which requires strong prior information and user interaction. No work has been devoted into direct estimation of RWT from cardiac MR images due to the diverse shapes and structures for various subjects and cardiac diseases, as well as the complex regional deformation of LV myocardium during the systole and diastole phases of the cardiac cycle. In this paper, we present a newly proposed Residual Recurrent Neural Network (ResRNN) that fully leverages the spatial and temporal dynamics of LV myocardium to achieve accurate frame-wise RWT estimation. Our ResRNN comprises two paths: 1) a feed forward convolution neural network (CNN) for effective and robust CNN embedding learning of various cardiac images and preliminary estimation of RWT from each frame itself independently, and 2) a recurrent neural network (RNN) for further improving the estimation by modeling spatial and temporal dynamics of LV myocardium. For the RNN path, we design for cardiac sequences a Circle-RNN to eliminate the effect of null hidden input for the first time-step. Our ResRNN is capable of obtaining accurate estimation of cardiac RWT with Mean Absolute Error of 1.44mm (less than 1-pixel error) when validated on cardiac MR sequences of 145 subjects, evidencing its great potential in clinical cardiac function assessment.

show abstract

“…5,6,[16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31] But LV segmentation is still an open problem and is challenging due to the poor contrast between tissues around the epicardium and image artifacts, such as intensity inhomogeneities in cine cardiac magnetic resonance (CMR) images. 6,12,19,[31][32][33][34][35][36][37] Traditional segmentation techniques, such as thresholding and k-means clustering, have been used to segment the LV from cardiac MR images. [38][39][40][41] Active shape modeling and active appearance modeling have also become prominent tools in LV segmentation.…”

Section: Introductionmentioning

confidence: 99%

Simultaneous extraction of endocardial and epicardial contours of the left ventricle by distance regularized level sets

et al. 2016

View full text Add to dashboard Cite

show abstract

Direct and Simultaneous Four-Chamber Volume Estimation by Multi-Output Regression

Cited by 36 publications

References 10 publications

Direct Multitype Cardiac Indices Estimation via Joint Representation and Regression Learning

Direct Multitype Cardiac Indices Estimation via Joint Representation and Regression Learning

Direct Estimation of Regional Wall Thicknesses via Residual Recurrent Neural Network

Simultaneous extraction of endocardial and epicardial contours of the left ventricle by distance regularized level sets

Contact Info

Product

Resources

About