High quality real-time Image-to-Mesh conversion for finite element simulations

Foteinos, Panagiotis A.; Chrisochoides, Nikos

doi:10.1016/j.jpdc.2013.11.002

Cited by 33 publications

(35 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In general, the timeconsuming and laborious nature of model construction makes the CFD or fluid structure interaction (FSI) simulations difficult for the large number of samples required for statistical comparisons. To reduce the time involved in generating solid models and improve the process efficiency, we are currently developing an image to mesh technique (I2M) to obtain finite element grids directly from segmented medical images [55][56][57]. One of the drawbacks of using the meshing tools in the commercial package ADINA, for example, is that it is restricted to working with Delaunay triangulation which introduces slivers in the finite element meshes when discretizing complex geometries involving a huge transition in edge sizes.…”

Section: Cfd Model Constructionmentioning

confidence: 99%

Hemodynamics of Flow Diverters

Dholakia

Sadasivan

Fiorella

et al. 2017

Journal of Biomechanical Engineering

View full text Add to dashboard Cite

Cerebral aneurysms are pathological focal evaginations of the arterial wall at and around the junctions of the circle of Willis. Their tenuous walls predispose aneurysms to leak or rupture leading to hemorrhagic strokes with high morbidity and mortality rates. The endovascular treatment of cerebral aneurysms currently includes the implantation of fine-mesh stents, called flow diverters, within the parent artery bearing the aneurysm. By mitigating flow velocities within the aneurysmal sac, the devices preferentially induce thrombus formation in the aneurysm within hours to days. In response to the foreign implant, an endothelialized arterial layer covers the luminal surface of the device over a period of days to months. Organization of the intraneurysmal thrombus leads to resorption and shrinkage of the aneurysm wall and contents, eventually leading to beneficial remodeling of the pathological site to a near-physiological state. The devices' primary function of reducing flow activity within aneurysms is corollary to their mesh structure. Complete specification of the device mesh structure, or alternately device permeability, necessarily involves the quantification of two variables commonly used to characterize porous media-mesh porosity and mesh pore density. We evaluated the flow alteration induced by five commercial neurovascular devices of varying porosity and pore density (stents: Neuroform, Enterprise, and LVIS; flow diverters: Pipeline and FRED) in an idealized sidewall aneurysm model. As can be expected in such a model, all devices substantially reduced intraneurysmal kinetic energy as compared to the nonstented case with the coarse-mesh stents inducing a 65-80% reduction whereas the fine-mesh flow diverters induced a near-complete flow stagnation (∼98% reduction). We also note a trend toward greater device efficacy (lower intraneurysmal flow) with decreasing device porosity and increasing device pore density. Several such flow studies have been and are being conducted in idealized as well as patient-derived geometries with the overarching goals of improving device design, facilitating treatment planning (what is the optimal device for a specific aneurysm), and predicting treatment outcome (will a specific aneurysm treated with a specific device successfully occlude over the long term). While the results are generally encouraging, there is poor standardization of study variables between different research groups, and any consensus will only be reached after standardized studies are conducted on collectively large datasets. Biochemical variables may have to be incorporated into these studies to maximize predictive values.

show abstract

Section: Cfd Model Constructionmentioning

confidence: 99%

Hemodynamics of Flow Diverters

Dholakia

Sadasivan

Fiorella

et al. 2017

Journal of Biomechanical Engineering

View full text Add to dashboard Cite

show abstract

“…The experimental evaluation was performed on Blacklight, a cache-3 coherent NUMA shared memory machine in the Pittsburgh Supercomputing Center. We observed a weak scaling efficiency of more than 67% on 192 cores, compared to only about 30% of that of PODM [8,9]. To the best of our knowledge, the LAPD so far delivered the best scalability result for three dimensional isosurface based parallel Delaunay mesh generation algorithms running on NUMA DSM supercomputers with quality and fidelity guaranteed.…”

Section: Introductionmentioning

confidence: 84%

“…The previous parallel mesh generation method, PODM [8,9], is a tightlycoupled parallel optimistic Delaunay mesh generation algorithm. PODM suffers from communication overhead caused by a large number of remote memory accesses, and its performance deteriorates for a core count beyond 144 because of the network congestion caused by the communication among threads.…”

Section: Background and Related Workmentioning

confidence: 99%

“…Delaunay mesh refinement is one of the most popular techniques for generating triangular and tetrahedral meshes for use in finite element analysis and interpolation in various numeric computing areas because it can mathematically guarantee the quality of the mesh [1,2,3,4]. The parallelization of Delaunay mesh refinement codes can be achieved by inserting multiple points simultaneously [5,6,7,8,9,10]. The mesh generation of the ultra-high-resolution three dimensional images, such as BigBrain [11], brings new challenges to parallel mesh generation considering the required memory space and the execution time [12].…”

Section: Introductionmentioning

confidence: 99%

“…In this paper, we conduct an application-level exploration that will help design general localityaware mesh generation algorithms and supporting run-time systems. In the previous work [8,9], we implemented a parallel Image-to-Mesh (I2M) refinement algorithm, Parallel Optimistic Delaunay Mesh generator (PODM), which works on multi-label segmented three dimensional images. The algorithm works well for a low core count (less than 128 cores).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Two-level locality-aware parallel Delaunay image-to-mesh conversion

2016

Self Cite

View full text Add to dashboard Cite

In this paper, we propose a three dimensional two-level Locality-Aware Parallel Delaunay image-to-mesh conversion algorithm (LAPD). The algorithm exploits two levels of parallelism at different granularities: coarse-grain parallelism at the region level (which is mapped to a node with multiple cores) and mediumgrain parallelism at the cavity level (which is mapped to a single core). We employ a data locality-aware mesh refinement process to reduce the latency caused by the remote memory access. We evaluated LAPD on Blacklight, a cache-coherent NUMA distributed shared memory (DSM) machine in the Pittsburgh Supercomputing Center, and observed a weak scaling efficiency of almost 70% for roughly 200 cores, compared to only 30% for the previous algorithm, Parallel Optimistic Mesh Generation algorithm (PODM). To the best of our knowledge, LAPD exhibits the best scalability for parallel Delaunay mesh generation algorithms running on NUMA DSM supercomputers.

show abstract

Non-rigid Registration of Brain MR Images for Image Guided Neurosurgery Using Cloud Computing

Evangeline

Anandhakumar

2019

Computer Aided Intervention and Diagnostics in Clinical and Medical Images

View full text Add to dashboard Cite

High quality real-time Image-to-Mesh conversion for finite element simulations

Cited by 33 publications

References 33 publications

Hemodynamics of Flow Diverters

Hemodynamics of Flow Diverters

Two-level locality-aware parallel Delaunay image-to-mesh conversion

Non-rigid Registration of Brain MR Images for Image Guided Neurosurgery Using Cloud Computing

Contact Info

Product

Resources

About