CNN-Based Cardiac Motion Extraction to Generate Deformable Geometric Left Ventricle Myocardial Models from Cine MRI Roshan Reddy Upendra 1(B ) , Brian Jamison Wentz 3,5 , Richard Simon 2 , Suzanne M. Shontz 3,4,5 , and Cristian A. Linte 1,2 1 Center for Imaging Science, Rochester Institute of Technology, Rochester, NY, USA ru6928@rit.edu 2 Biomedical Engineering, Rochester Institute of Technology, Rochester, NY, USA 3 Bioengineering Program, University of Kansas, Lawrence, KS, USA 4 Electrical Engineering and Computer Science, University of Kansas, Lawrence, KS, USA 5 Information and Telecommunication Center, University of Kansas, Lawrence, KS, USA Abstract. Patient-specific left ventricle (LV) myocardial models have the potential to be used in a variety of clinical scenarios for improved diagnosis and treatment plans. Cine cardiac magnetic resonance (MR) imaging provides high resolution images to reconstruct patient-specific geometric models of the LV myocardium. With the advent of deep learn- ing, accurate segmentation of cardiac chambers from cine cardiac MR images and unsupervised learning for image registration for cardiac motion estimation on a large number of image datasets is attainable. Here, we propose a deep leaning-based framework for the development of patient-specific geometric models of LV myocardium from cine cardiac MR images, using the Automated Cardiac Diagnosis Challenge (ACDC) dataset. We use the deformation field estimated from the VoxelMorph- based convolutional neural network (CNN) to propagate the isosurface mesh and volume mesh of the end-diastole (ED) frame to the subsequent frames of the cardiac cycle. We assess the CNN-based propagated mod- els against segmented models at each cardiac phase, as well as models propagated using another traditional nonrigid image registration tech- nique. Additionally, we generate dynamic LV myocardial volume meshes at all phases of the cardiac cycle using the log barrier-based mesh warp- ing (LBWARP) method and compare them with the CNN-propagated volume meshes. Keywords: Patient-specific modeling · Deep learning · Image registration · Cine Cardiac MRI · Mesh warping 1 Introduction To reduce the morbidity and mortality associated with cardiovascular diseases (CVDs) [3], and to improve their treatment, it is crucial to detect and predict the c  Springer Nature Switzerland AG 2021 D. B. Ennis et al. (Eds.): FIMH 2021, LNCS 12738, pp. 253–263, 2021. https://doi.org/10.1007/978-3-030-78710-3_25