No abstract

No abstract

No abstract

Cardiac mechanics tools can be used to enhance medical diagnosis and treatment, and assessment of risk of cardiovascular diseases. Still, the computational cost to solve cardiac models restricts their use for online applications and routine clinical practice. This work presents a surrogate model obtained by training a set of Siamese networks over a physiological representation of the left ventricle. Our model allows us to modify the geometry, loading conditions, and material properties without needing of retraining. Additionally, we propose the novel concept of intrinsic domain that improves the accuracy of the network predictions by one order of magnitude. The neural networks were trained and tested with numerical predictions from a previously published finite element model of the left ventricle. Different loading conditions, material properties and geometrical definitions of the domain were simulated by the model leading to a dataset of 5, 670 cases. In terms of accuracy and performance, the surrogate model approximates the displacement field of the finite element model with an error of 4.4 ± 2.9% (with respect to the L 2-norm of the true displacement field) across all cases while performing computations 62 times faster. Hence, the trained model is capable of computing a passive cardiac filling of the chamber at 10 different time points in just ∼ 0.7 s. These outcomes prove usability of training surrogate models for efficient simulations to facilitate the use of complex mechanical models in clinical practice for therapeutic planning and online diagnosis.

This study investigates the use of multiple gravity‐loaded configurations of a soft heterogeneous body for the identification of its mechanical properties. It is largely motivated by the need for in vivo, non‐invasive determination of the mechanical properties of biological tissue, such as the breast, for the purposes of tracking and predicting tumor locations. In order to validate this approach, experiments were performed on a heterogeneous two‐layered silicone gel cantilever beam, laser scanned in eight different orientations under gravity loading. A finite element model of the beam was constructed and analyzed using non‐linear finite elasticity and a neo‐Hookean stress–strain relationship. The constitutive parameters representing the stiffness of each layer were estimated, using non‐linear optimization techniques, to best fit the laser‐scanned data. Different subsets of the experiments were used for training (parameter fitting), with the remaining experiments being used for validation. In both experimental sets, a cross‐validation study was performed in order to compare and assess the predictive power of the identified parameters. Several determinability criteria were used to assess the objective function in the neighborhood of the identified minimum. We found that individual experiments were not sufficient for reliable parameter identification of heterogeneous soft bodies under gravity loading, but that a small number of gravity‐loaded configurations were sufficient to reliably estimate the parameters within the scale of experimental errors. We also discuss the practical and numerical issues related to both single and multi‐experiment parameter estimation. Copyright © 2011 John Wiley & Sons, Ltd.