A Method of Detecting Surface Mesh Quality Based on Deep Learning

AbstractGraphein is a python library for constructing graph and surface-mesh representations of protein structures for computational analysis. The library interfaces with popular geometric deep learning libraries: DGL, PyTorch Geometric and PyTorch3D. Geometric deep learning is emerging as a popular methodology in computational structural biology. As feature engineering is a vital step in a machine learning project, the library is designed to be highly flexible, allowing the user to parameterise the graph construction, scaleable to facilitate working with large protein complexes, and containing useful pre-processing tools for preparing experimental structure files. Graphein is also designed to facilitate network-based and graph-theoretic analyses of protein structures in a high-throughput manner. As example workflows, we make available two new protein structure-related datasets, previously unused by the geometric deep learning community.Availability and implementationGraphein is written in python. Source code, example usage and datasets, and documentation are made freely available under a MIT License at the following URL: https://github.com/a-r-j/graphein

Download Full-text

Combination Of Handcrafted And Deep Learning-Based Features For 3d Mesh Quality Assessment

2020 IEEE International Conference on Image Processing (ICIP) ◽

10.1109/icip40778.2020.9191092 ◽

2020 ◽

Author(s):

Ilyass Abouelaziz ◽

Aladine Chetouani ◽

Mohammed El Hassouni ◽

Longin Jan Latecki ◽

Hocine Cherifi

Keyword(s):

Deep Learning ◽

Quality Assessment ◽

Mesh Quality ◽

3D Mesh

Download Full-text

Surface mesh quality improvement by weighted inscribed circle method

Advances in Engineering Software ◽

10.1016/j.advengsoft.2019.102693 ◽

2019 ◽

Vol 136 ◽

pp. 102693

Author(s):

Wei Peng ◽

Weixi Ji

Keyword(s):

Quality Improvement ◽

Circle Method ◽

Surface Mesh ◽

Mesh Quality ◽

Mesh Quality Improvement

Download Full-text

Optimization of surface mesh quality with feature preserving

The proceedings of the JSME annual meeting ◽

10.1299/jsmemecjo.2004.6.0_87 ◽

2004 ◽

Vol 2004.6 (0) ◽

pp. 87-88

Author(s):

Irina Semenova ◽

Ichiro HAGIWARA

Keyword(s):

Surface Mesh ◽

Mesh Quality ◽

Feature Preserving

Download Full-text

Finite triangular surface mesh simplification with geometrical feature recognition

Proceedings of the Institution of Mechanical Engineers Part C Journal of Mechanical Engineering Science ◽

10.1243/09544062jmes1506 ◽

2009 ◽

Vol 223 (11) ◽

pp. 2627-2636 ◽

Cited By ~ 2

Author(s):

Y Zhang ◽

H Wang ◽

H Zhou ◽

J Li

Keyword(s):

Feature Recognition ◽

Injection Pressure ◽

Fem Analysis ◽

Mesh Simplification ◽

Surface Mesh ◽

Mesh Quality ◽

Large Curvature ◽

Geometrical Feature ◽

Aspect Ratios ◽

Triangular Surface Mesh

The quantity and quality of the mesh elements are both significant factors for guaranteeing computational precision in the finite-element method (FEM). In most mesh simplification algorithms, the geometric error was thought to be the most important issue. However, the mesh quality was rarely taken into consideration. In this article, a finite triangular surface mesh simplification algorithm is proposed, in which the vertex dispersion is introduced to represent the local geometrical feature, and then either edge collapse or face collapse is carried out consequently. The aspect of the newly created face is taken into account and the position of the newly created vertex is obtained by solving an over-determined system of linear equations with respect to the aspect ratios of the newly created faces; thereupon the simplified mesh quality is improved. To obtain a further simplification and to reduce the errors in the succeeding FEM analysis, surface fitting is adopted on the surfaces with large curvature. Simplification cases had been performed in comparison with the quadric error metric method, and the results show that the definition of the local mesh density is more reasonable for the FEM analysis with the same simplification ratio while the present simplification algorithm is employed. Moreover, the mesh quality can be greatly improved on the surface with large curvature. A set of FEM numerical experiments of polymer injection moulding simulation had also been performed to determine the effect of the presented simplification algorithm on the FEM analysis. The numerical results show that the error of the injection pressure can be limited within 4 per cent, while the simplification percentage reaches 75 per cent.

Download Full-text