Rotation Invariant Graph Neural Network for 3D Point Clouds
In this paper we propose a novel rotation normalization technique for point cloud processing using an oriented bounding box. We use this method to create a point cloud annotation tool for part segmentation on real camera data. Custom data sets are used to train our network for classification and par...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-03-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/15/5/1437 |
| _version_ | 1797614464645726208 |
|---|---|
| author | Alexandru Pop Victor Domșa Levente Tamas |
| author_facet | Alexandru Pop Victor Domșa Levente Tamas |
| author_sort | Alexandru Pop |
| collection | DOAJ |
| description | In this paper we propose a novel rotation normalization technique for point cloud processing using an oriented bounding box. We use this method to create a point cloud annotation tool for part segmentation on real camera data. Custom data sets are used to train our network for classification and part segmentation tasks. Successful deployment is completed on an embedded device with limited processing power. A comparison is made with other rotation-invariant features in noisy synthetic datasets. Our method offers more auxiliary information related to the dimension, position, and orientation of the object than previous methods while performing at a similar level. |
| first_indexed | 2024-03-11T07:11:46Z |
| format | Article |
| id | doaj.art-af3cbedba79d479cac3f9c692c00f033 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-11T07:11:46Z |
| publishDate | 2023-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-af3cbedba79d479cac3f9c692c00f0332023-11-17T08:33:02ZengMDPI AGRemote Sensing2072-42922023-03-01155143710.3390/rs15051437Rotation Invariant Graph Neural Network for 3D Point CloudsAlexandru Pop0Victor Domșa1Levente Tamas2Automation Department, Technical University of Cluj-Napoca, 400114 Cluj-Napoca, RomaniaAutomation Department, Technical University of Cluj-Napoca, 400114 Cluj-Napoca, RomaniaAutomation Department, Technical University of Cluj-Napoca, 400114 Cluj-Napoca, RomaniaIn this paper we propose a novel rotation normalization technique for point cloud processing using an oriented bounding box. We use this method to create a point cloud annotation tool for part segmentation on real camera data. Custom data sets are used to train our network for classification and part segmentation tasks. Successful deployment is completed on an embedded device with limited processing power. A comparison is made with other rotation-invariant features in noisy synthetic datasets. Our method offers more auxiliary information related to the dimension, position, and orientation of the object than previous methods while performing at a similar level.https://www.mdpi.com/2072-4292/15/5/1437computer visionobject part segmentationclassification |
| spellingShingle | Alexandru Pop Victor Domșa Levente Tamas Rotation Invariant Graph Neural Network for 3D Point Clouds Remote Sensing computer vision object part segmentation classification |
| title | Rotation Invariant Graph Neural Network for 3D Point Clouds |
| title_full | Rotation Invariant Graph Neural Network for 3D Point Clouds |
| title_fullStr | Rotation Invariant Graph Neural Network for 3D Point Clouds |
| title_full_unstemmed | Rotation Invariant Graph Neural Network for 3D Point Clouds |
| title_short | Rotation Invariant Graph Neural Network for 3D Point Clouds |
| title_sort | rotation invariant graph neural network for 3d point clouds |
| topic | computer vision object part segmentation classification |
| url | https://www.mdpi.com/2072-4292/15/5/1437 |
| work_keys_str_mv | AT alexandrupop rotationinvariantgraphneuralnetworkfor3dpointclouds AT victordomsa rotationinvariantgraphneuralnetworkfor3dpointclouds AT leventetamas rotationinvariantgraphneuralnetworkfor3dpointclouds |