A GPU-Based Framework for Generating Implicit Datasets of Voxelized Polygonal Models for the Training of 3D Convolutional Neural Networks
In this paper we present an efficient GPU-based framework to dynamically perform the voxelization of polygonal models for training 3D convolutional neural networks. It is designed to manage the dataset augmentation by using efficient geometric transformations and random vertex displacements in GPU....
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8955843/ |
| _version_ | 1818435752589000704 |
|---|---|
| author | Carlos J. Ogayar-Anguita Antonio J. Rueda-Ruiz Rafael J. Segura-Sanchez Miguel Diaz-Medina Angel L. Garcia-Fernandez |
| author_facet | Carlos J. Ogayar-Anguita Antonio J. Rueda-Ruiz Rafael J. Segura-Sanchez Miguel Diaz-Medina Angel L. Garcia-Fernandez |
| author_sort | Carlos J. Ogayar-Anguita |
| collection | DOAJ |
| description | In this paper we present an efficient GPU-based framework to dynamically perform the voxelization of polygonal models for training 3D convolutional neural networks. It is designed to manage the dataset augmentation by using efficient geometric transformations and random vertex displacements in GPU. With the proposed system, every voxelization is carried out on-the-fly for directly feeding the network. The computing performance with this approach is much better than with the standard method, that carries out every voxelization of each model separately and has much higher data processing overhead. The core voxelization algorithm works by using the standard rendering pipeline of the GPU. It generates binary voxels for both the interior and the surface of the models. Moreover, it is simple, concise and very compatible with commodity hardware, as it neither uses complex data structures nor needs vendor-specific hardware or additional dependencies. This framework dramatically reduces the input/output operations, and completely eliminates the storage footprint of the voxelization dataset, managing it as an implicit dataset. |
| first_indexed | 2024-12-14T16:57:53Z |
| format | Article |
| id | doaj.art-f0b6dc5c842449b782a9abdd8fe79eed |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-14T16:57:53Z |
| publishDate | 2020-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-f0b6dc5c842449b782a9abdd8fe79eed2022-12-21T22:53:56ZengIEEEIEEE Access2169-35362020-01-018126751268710.1109/ACCESS.2020.29656248955843A GPU-Based Framework for Generating Implicit Datasets of Voxelized Polygonal Models for the Training of 3D Convolutional Neural NetworksCarlos J. Ogayar-Anguita0https://orcid.org/0000-0003-0958-990XAntonio J. Rueda-Ruiz1https://orcid.org/0000-0001-7692-454XRafael J. Segura-Sanchez2https://orcid.org/0000-0002-3075-6963Miguel Diaz-Medina3https://orcid.org/0000-0003-2577-323XAngel L. Garcia-Fernandez4https://orcid.org/0000-0002-8183-7130Department of Computer Science, EPS Jaén, University of Jaén, Jaén, SpainDepartment of Computer Science, EPS Jaén, University of Jaén, Jaén, SpainDepartment of Computer Science, EPS Jaén, University of Jaén, Jaén, SpainDepartment of Computer Science, EPS Jaén, University of Jaén, Jaén, SpainDepartment of Computer Science, EPS Jaén, University of Jaén, Jaén, SpainIn this paper we present an efficient GPU-based framework to dynamically perform the voxelization of polygonal models for training 3D convolutional neural networks. It is designed to manage the dataset augmentation by using efficient geometric transformations and random vertex displacements in GPU. With the proposed system, every voxelization is carried out on-the-fly for directly feeding the network. The computing performance with this approach is much better than with the standard method, that carries out every voxelization of each model separately and has much higher data processing overhead. The core voxelization algorithm works by using the standard rendering pipeline of the GPU. It generates binary voxels for both the interior and the surface of the models. Moreover, it is simple, concise and very compatible with commodity hardware, as it neither uses complex data structures nor needs vendor-specific hardware or additional dependencies. This framework dramatically reduces the input/output operations, and completely eliminates the storage footprint of the voxelization dataset, managing it as an implicit dataset.https://ieeexplore.ieee.org/document/8955843/VoxelizationB-repboundary representationpolygonal meshesconvolutional neural network3D-CNN |
| spellingShingle | Carlos J. Ogayar-Anguita Antonio J. Rueda-Ruiz Rafael J. Segura-Sanchez Miguel Diaz-Medina Angel L. Garcia-Fernandez A GPU-Based Framework for Generating Implicit Datasets of Voxelized Polygonal Models for the Training of 3D Convolutional Neural Networks IEEE Access Voxelization B-rep boundary representation polygonal meshes convolutional neural network 3D-CNN |
| title | A GPU-Based Framework for Generating Implicit Datasets of Voxelized Polygonal Models for the Training of 3D Convolutional Neural Networks |
| title_full | A GPU-Based Framework for Generating Implicit Datasets of Voxelized Polygonal Models for the Training of 3D Convolutional Neural Networks |
| title_fullStr | A GPU-Based Framework for Generating Implicit Datasets of Voxelized Polygonal Models for the Training of 3D Convolutional Neural Networks |
| title_full_unstemmed | A GPU-Based Framework for Generating Implicit Datasets of Voxelized Polygonal Models for the Training of 3D Convolutional Neural Networks |
| title_short | A GPU-Based Framework for Generating Implicit Datasets of Voxelized Polygonal Models for the Training of 3D Convolutional Neural Networks |
| title_sort | gpu based framework for generating implicit datasets of voxelized polygonal models for the training of 3d convolutional neural networks |
| topic | Voxelization B-rep boundary representation polygonal meshes convolutional neural network 3D-CNN |
| url | https://ieeexplore.ieee.org/document/8955843/ |
| work_keys_str_mv | AT carlosjogayaranguita agpubasedframeworkforgeneratingimplicitdatasetsofvoxelizedpolygonalmodelsforthetrainingof3dconvolutionalneuralnetworks AT antoniojruedaruiz agpubasedframeworkforgeneratingimplicitdatasetsofvoxelizedpolygonalmodelsforthetrainingof3dconvolutionalneuralnetworks AT rafaeljsegurasanchez agpubasedframeworkforgeneratingimplicitdatasetsofvoxelizedpolygonalmodelsforthetrainingof3dconvolutionalneuralnetworks AT migueldiazmedina agpubasedframeworkforgeneratingimplicitdatasetsofvoxelizedpolygonalmodelsforthetrainingof3dconvolutionalneuralnetworks AT angellgarciafernandez agpubasedframeworkforgeneratingimplicitdatasetsofvoxelizedpolygonalmodelsforthetrainingof3dconvolutionalneuralnetworks AT carlosjogayaranguita gpubasedframeworkforgeneratingimplicitdatasetsofvoxelizedpolygonalmodelsforthetrainingof3dconvolutionalneuralnetworks AT antoniojruedaruiz gpubasedframeworkforgeneratingimplicitdatasetsofvoxelizedpolygonalmodelsforthetrainingof3dconvolutionalneuralnetworks AT rafaeljsegurasanchez gpubasedframeworkforgeneratingimplicitdatasetsofvoxelizedpolygonalmodelsforthetrainingof3dconvolutionalneuralnetworks AT migueldiazmedina gpubasedframeworkforgeneratingimplicitdatasetsofvoxelizedpolygonalmodelsforthetrainingof3dconvolutionalneuralnetworks AT angellgarciafernandez gpubasedframeworkforgeneratingimplicitdatasetsofvoxelizedpolygonalmodelsforthetrainingof3dconvolutionalneuralnetworks |