Exploring local delaunay graph based neural network for 3D object detection

Show simple item record

dc.contributor.advisor Poovvancheri, Jiju
dc.creator Pandey, Bivash
dc.date.accessioned 2022-05-06T19:52:56Z
dc.date.available 2022-05-06T19:52:56Z
dc.date.issued 2022-04-26
dc.identifier.uri http://library2.smu.ca/xmlui/handle/01/30912
dc.description 1 online resource (v, 41 pages) : colour illustrations, colour charts, graphs
dc.description Includes abstract.
dc.description Includes bibliographical references (pages 36-41).
dc.description.abstract Three-dimensional object detectors are essential components of the perception sys tems of autonomous robots. Most of the 3D object detection methods are point-based, voxel-based, or point-voxel-based (uses both point-based and voxel-based approaches). Recently, few works have used graph neural networks for 3D object detection and achieved promising results. Representation of a point cloud as a graph preserves the irregularity and removes the cost associated with the transformation into voxel grids or projection into a bird’s eye view. In this thesis, we study the effects of point cloud encoding using local Delaunay graphs by embedding it in an existing graph neural network (Point-GNN). We perform experiments on the KITTI benchmark and find that graph-based methods achieve higher or comparable accuracy. Our results are comparable with most of the state-of-the-art methods. Further, our study indicates that the selection of the right type of proximity graph representation is crucial for real-time 3D object detection. en_CA
dc.description.provenance Submitted by Greg Hilliard (greg.hilliard@smu.ca) on 2022-05-06T19:52:56Z No. of bitstreams: 1 Pandey_Bivash_Honours_2022.pdf: 14290021 bytes, checksum: 2d7b048ef3f43b456d1cfc962ed73ea3 (MD5) en
dc.description.provenance Made available in DSpace on 2022-05-06T19:52:56Z (GMT). No. of bitstreams: 1 Pandey_Bivash_Honours_2022.pdf: 14290021 bytes, checksum: 2d7b048ef3f43b456d1cfc962ed73ea3 (MD5) Previous issue date: 2022-04-26 en
dc.language.iso en en_CA
dc.publisher Halifax, N.S. : Saint Mary's University
dc.title Exploring local delaunay graph based neural network for 3D object detection en_CA
dc.type Text en_CA
thesis.degree.name Bachelor of Science (Honours Computing Science)
thesis.degree.level Undergraduate
thesis.degree.discipline Mathematics and Computing Science
thesis.degree.grantor Saint Mary's University (Halifax, N.S.)
 Find Full text

Files in this item

 
 

This item appears in the following Collection(s)

Show simple item record

Search DSpace


Browse

My Account