EFFICIENT TRAINING OF SEMANTIC POINT CLOUD SEGMENTATION VIA ACTIVE LEARNING
With the development of LiDAR and photogrammetric techniques, more and more point clouds are available with high density and in large areas. Point cloud interpretation is an important step before many real applications like 3D city modelling. Many supervised machine learning techniques have been ada...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2020-08-01
|
| Series: | ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| Online Access: | https://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/V-2-2020/243/2020/isprs-annals-V-2-2020-243-2020.pdf |
| _version_ | 1818475164449374208 |
|---|---|
| author | Y. Lin G. Vosselman Y. Cao M. Y. Yang |
| author_facet | Y. Lin G. Vosselman Y. Cao M. Y. Yang |
| author_sort | Y. Lin |
| collection | DOAJ |
| description | With the development of LiDAR and photogrammetric techniques, more and more point clouds are available with high density and in large areas. Point cloud interpretation is an important step before many real applications like 3D city modelling. Many supervised machine learning techniques have been adapted to semantic point cloud segmentation, aiming to automatically label point clouds. Current deep learning methods have shown their potentials to produce high accuracy in semantic point cloud segmentation tasks. However, these supervised methods require a large amount of labelled data for proper model performance and good generalization. In practice, manual labelling of point clouds is very expensive and time-consuming. Active learning can iteratively select unlabelled samples for manual annotation based on current statistical models and then update the labelled data pool for next model training. In order to effectively label point clouds, we proposed a segment based active learning strategy to assess the informativeness of samples. Here, the proposed strategy uses 40% of the whole training dataset to achieve a mean IoU of 75.2% which is 99.1% of the accuracy in mIoU obtained from the model trained on the full dataset, while the baseline method using same amount of data only reaches 69.6% in mIoU corresponding to 90.9% of the accuracy in mIoU obtained from the model trained on the full dataset. |
| first_indexed | 2024-12-10T09:09:27Z |
| format | Article |
| id | doaj.art-2022738550204a4f958b9378302fa2f8 |
| institution | Directory Open Access Journal |
| issn | 2194-9042 2194-9050 |
| language | English |
| last_indexed | 2024-12-10T09:09:27Z |
| publishDate | 2020-08-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| spelling | doaj.art-2022738550204a4f958b9378302fa2f82022-12-22T01:55:04ZengCopernicus PublicationsISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences2194-90422194-90502020-08-01V-2-202024325010.5194/isprs-annals-V-2-2020-243-2020EFFICIENT TRAINING OF SEMANTIC POINT CLOUD SEGMENTATION VIA ACTIVE LEARNINGY. Lin0G. Vosselman1Y. Cao2M. Y. Yang3Dept. of Earth Observation Science, Faculty ITC, University of Twente, Enschede, The NetherlandsDept. of Earth Observation Science, Faculty ITC, University of Twente, Enschede, The NetherlandsState Key Laboratory of Fluid Power and Mechatronic Systems, School of Mechanical Engineering, Zhejiang University, Hangzhou, ChinaDept. of Earth Observation Science, Faculty ITC, University of Twente, Enschede, The NetherlandsWith the development of LiDAR and photogrammetric techniques, more and more point clouds are available with high density and in large areas. Point cloud interpretation is an important step before many real applications like 3D city modelling. Many supervised machine learning techniques have been adapted to semantic point cloud segmentation, aiming to automatically label point clouds. Current deep learning methods have shown their potentials to produce high accuracy in semantic point cloud segmentation tasks. However, these supervised methods require a large amount of labelled data for proper model performance and good generalization. In practice, manual labelling of point clouds is very expensive and time-consuming. Active learning can iteratively select unlabelled samples for manual annotation based on current statistical models and then update the labelled data pool for next model training. In order to effectively label point clouds, we proposed a segment based active learning strategy to assess the informativeness of samples. Here, the proposed strategy uses 40% of the whole training dataset to achieve a mean IoU of 75.2% which is 99.1% of the accuracy in mIoU obtained from the model trained on the full dataset, while the baseline method using same amount of data only reaches 69.6% in mIoU corresponding to 90.9% of the accuracy in mIoU obtained from the model trained on the full dataset.https://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/V-2-2020/243/2020/isprs-annals-V-2-2020-243-2020.pdf |
| spellingShingle | Y. Lin G. Vosselman Y. Cao M. Y. Yang EFFICIENT TRAINING OF SEMANTIC POINT CLOUD SEGMENTATION VIA ACTIVE LEARNING ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| title | EFFICIENT TRAINING OF SEMANTIC POINT CLOUD SEGMENTATION VIA ACTIVE LEARNING |
| title_full | EFFICIENT TRAINING OF SEMANTIC POINT CLOUD SEGMENTATION VIA ACTIVE LEARNING |
| title_fullStr | EFFICIENT TRAINING OF SEMANTIC POINT CLOUD SEGMENTATION VIA ACTIVE LEARNING |
| title_full_unstemmed | EFFICIENT TRAINING OF SEMANTIC POINT CLOUD SEGMENTATION VIA ACTIVE LEARNING |
| title_short | EFFICIENT TRAINING OF SEMANTIC POINT CLOUD SEGMENTATION VIA ACTIVE LEARNING |
| title_sort | efficient training of semantic point cloud segmentation via active learning |
| url | https://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/V-2-2020/243/2020/isprs-annals-V-2-2020-243-2020.pdf |
| work_keys_str_mv | AT ylin efficienttrainingofsemanticpointcloudsegmentationviaactivelearning AT gvosselman efficienttrainingofsemanticpointcloudsegmentationviaactivelearning AT ycao efficienttrainingofsemanticpointcloudsegmentationviaactivelearning AT myyang efficienttrainingofsemanticpointcloudsegmentationviaactivelearning |