USING EDGECONV TO IMPROVE 3D OBJECT DETECTION FROM RGB-D DATA

USING EDGECONV TO IMPROVE 3D OBJECT DETECTION FROM RGB-D DATA

In this paper, we proposed a novel 3D deep learning model for object localization and object bounding boxes estimation. To increase the detection efficiency of small objects in the large scale scenes, the local neighbourhood geometric structure information of objects has been taken into the Edgeconv...

Full description

Bibliographic Details
Main Authors: W. Lin, Y. Chen, C. Wang, J. Li
Format: Article
Language:English
Published: Copernicus Publications 2019-06-01
Series:The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Online Access:https://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XLII-2-W13/835/2019/isprs-archives-XLII-2-W13-835-2019.pdf
id doaj-e3b13073d68949ac827d8f08a958ad87
record_format Article
spelling doaj-e3b13073d68949ac827d8f08a958ad872020-11-25T01:30:20ZengCopernicus PublicationsThe International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences1682-17502194-90342019-06-01XLII-2-W1383583910.5194/isprs-archives-XLII-2-W13-835-2019USING EDGECONV TO IMPROVE 3D OBJECT DETECTION FROM RGB-D DATAW. Lin0Y. Chen1C. Wang2J. Li3J. Li4Fujian Key Laboratory of Sensing and Computing, School of Informatics, Xiamen University, 422 Siming Road South, Xiamen 361005, ChinaFujian Key Laboratory of Sensing and Computing, School of Informatics, Xiamen University, 422 Siming Road South, Xiamen 361005, ChinaFujian Key Laboratory of Sensing and Computing, School of Informatics, Xiamen University, 422 Siming Road South, Xiamen 361005, ChinaFujian Key Laboratory of Sensing and Computing, School of Informatics, Xiamen University, 422 Siming Road South, Xiamen 361005, ChinaMobile Mapping Lab, Department of Geography and Environmental Management, University of Waterloo, Waterloo, ON N2L 3G1, CanadaIn this paper, we proposed a novel 3D deep learning model for object localization and object bounding boxes estimation. To increase the detection efficiency of small objects in the large scale scenes, the local neighbourhood geometric structure information of objects has been taken into the Edgeconv model, which can operate the original point clouds. We evaluated the 3D bounding box with high resolution in the RGB-D dataset and acquired stable effectiveness even under the sparse points and the strong occlusion. The experimental results indicate that our method achieved the higher mean average precision and better IOU of bounding boxes in SUN RGB-D dataset and KITTI benchmark.https://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XLII-2-W13/835/2019/isprs-archives-XLII-2-W13-835-2019.pdf
collection DOAJ
language English
format Article
sources DOAJ
author W. Lin
Y. Chen
C. Wang
J. Li
J. Li
spellingShingle W. Lin
Y. Chen
C. Wang
J. Li
J. Li
USING EDGECONV TO IMPROVE 3D OBJECT DETECTION FROM RGB-D DATA
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
author_facet W. Lin
Y. Chen
C. Wang
J. Li
J. Li
author_sort W. Lin
title USING EDGECONV TO IMPROVE 3D OBJECT DETECTION FROM RGB-D DATA
title_short USING EDGECONV TO IMPROVE 3D OBJECT DETECTION FROM RGB-D DATA
title_full USING EDGECONV TO IMPROVE 3D OBJECT DETECTION FROM RGB-D DATA
title_fullStr USING EDGECONV TO IMPROVE 3D OBJECT DETECTION FROM RGB-D DATA
title_full_unstemmed USING EDGECONV TO IMPROVE 3D OBJECT DETECTION FROM RGB-D DATA
title_sort using edgeconv to improve 3d object detection from rgb-d data
publisher Copernicus Publications
series The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
issn 1682-1750
2194-9034
publishDate 2019-06-01
description In this paper, we proposed a novel 3D deep learning model for object localization and object bounding boxes estimation. To increase the detection efficiency of small objects in the large scale scenes, the local neighbourhood geometric structure information of objects has been taken into the Edgeconv model, which can operate the original point clouds. We evaluated the 3D bounding box with high resolution in the RGB-D dataset and acquired stable effectiveness even under the sparse points and the strong occlusion. The experimental results indicate that our method achieved the higher mean average precision and better IOU of bounding boxes in SUN RGB-D dataset and KITTI benchmark.
url https://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XLII-2-W13/835/2019/isprs-archives-XLII-2-W13-835-2019.pdf
work_keys_str_mv AT wlin usingedgeconvtoimprove3dobjectdetectionfromrgbddata
AT ychen usingedgeconvtoimprove3dobjectdetectionfromrgbddata
AT cwang usingedgeconvtoimprove3dobjectdetectionfromrgbddata
AT jli usingedgeconvtoimprove3dobjectdetectionfromrgbddata
AT jli usingedgeconvtoimprove3dobjectdetectionfromrgbddata
_version_ 1725092039688716288

Similar Items