Please use this identifier to cite or link to this item:
https://hdl.handle.net/10316/114689
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Erabati, Gopi Krishna | - |
dc.contributor.author | Araújo, Helder | - |
dc.date.accessioned | 2024-04-05T07:38:13Z | - |
dc.date.available | 2024-04-05T07:38:13Z | - |
dc.date.issued | 2023 | - |
dc.identifier.uri | https://hdl.handle.net/10316/114689 | - |
dc.description.abstract | Inspired by recent advances in vision transformers for object detection, we propose Li3DeTr, an end-to-end LiDAR based 3D Detection Transformer for autonomous driving, that inputs LiDAR point clouds and regresses 3D bounding boxes. The LiDAR local and global features are encoded using sparse convolution and multi-scale deformable attention respectively. In the decoder head, firstly, in the novel Li3DeTr cross-attention block, we link the LiDAR global features to 3D predictions leveraging the sparse set of object queries learnt from the data. Secondly, the object query interactions are formulated using multi-head self-attention. Finally, the decoder layer is repeated Ldec number of times to refine the object queries. Inspired by DETR, we employ set-to-set loss to train the Li3DeTr network. Without bells and whistles, the Li3DeTr network achieves 61.3% mAP and 67.6% NDS surpassing the state-of-the-art methods with non-maximum suppression (NMS) on the nuScenes dataset and it also achieves competitive performance on the KITTI dataset. We also employ knowledge distillation (KD) using a teacher and student model that slightly improves the performance of our network. | pt |
dc.language.iso | eng | pt |
dc.publisher | IEEE | pt |
dc.relation | European Union’s H2020 MSCA-ITN-ACHIEVE with grant agreement No. 765866 | pt |
dc.relation | UIDB/00048/2020 | pt |
dc.relation | FCT Portugal PhD research grant with reference 2021.06219.BD | pt |
dc.rights | openAccess | pt |
dc.title | Li3DeTr: A LiDAR based 3D Detection Transformer | pt |
dc.type | article | - |
degois.publication.firstPage | 4239 | pt |
degois.publication.lastPage | 4248 | pt |
degois.publication.title | Proceedings - 2023 IEEE Winter Conference on Applications of Computer Vision, WACV 2023 | pt |
dc.peerreviewed | yes | pt |
dc.identifier.doi | 10.1109/WACV56688.2023.00423 | pt |
dc.date.embargo | 2023-01-01 | * |
uc.date.periodoEmbargo | 0 | pt |
item.fulltext | Com Texto completo | - |
item.openairecristype | http://purl.org/coar/resource_type/c_18cf | - |
item.languageiso639-1 | en | - |
item.openairetype | article | - |
item.cerifentitytype | Publications | - |
item.grantfulltext | open | - |
crisitem.project.grantno | INSTITUTE OF SYSTEMS AND ROBOTICS - ISR - COIMBRA | - |
crisitem.author.dept | ISR - Institute of Systems and Robotics | - |
crisitem.author.parentdept | University of Coimbra | - |
crisitem.author.researchunit | ISR - Institute of Systems and Robotics | - |
crisitem.author.parentresearchunit | University of Coimbra | - |
crisitem.author.orcid | 0000-0002-9544-424X | - |
Appears in Collections: | I&D ISR - Artigos em Revistas Internacionais |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Li3DeTr_A_LiDAR_based_3D_Detection_Transformer.pdf | 1.28 MB | Adobe PDF | View/Open |
Page view(s)
67
checked on Oct 16, 2024
Download(s)
43
checked on Oct 16, 2024
Google ScholarTM
Check
Altmetric
Altmetric
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.