EdgeNet: Semantic scene completion from a single RGB-D image
EdgeNet: Semantic scene completion from a single RGB-D image
Semantic scene completion is the task of predicting a complete 3D representation of volumetric occupancy with corresponding semantic labels for a scene from a single point of view. In this paper, we present EdgeNet, a new end-to-end neural network architecture that fuses information from depth and RGB, explicitly representing RGB edges in 3D space. Previous works on this task used either depth-only or depth with colour by projecting 2D semantic labels generated by a 2D segmentation network into the 3D volume, requiring a two step training process. Our EdgeNet representation encodes colour information in 3D space using edge detection and flipped truncated signed distance, which improves semantic completion scores especially in hard to detect classes. We achieved state-of-the-art scores on both synthetic and real datasets with a simpler and a more computationally efficient training pipeline than competing approaches.
503-510
Dourado, Aloisio
93f6e1e2-1fdd-4d33-8f08-b0df73cf5cb1
de Campos, Teofilo E.
dca09f0a-744c-4a15-a023-0a8f786c55d3
Kim, Hansung
2c7c135c-f00b-4409-acb2-85b3a9e8225f
Hilton, Adrian
12782a55-4c4d-4dfb-a690-62505f6665db
15 January 2021
Dourado, Aloisio
93f6e1e2-1fdd-4d33-8f08-b0df73cf5cb1
de Campos, Teofilo E.
dca09f0a-744c-4a15-a023-0a8f786c55d3
Kim, Hansung
2c7c135c-f00b-4409-acb2-85b3a9e8225f
Hilton, Adrian
12782a55-4c4d-4dfb-a690-62505f6665db
Dourado, Aloisio, de Campos, Teofilo E., Kim, Hansung and Hilton, Adrian
(2021)
EdgeNet: Semantic scene completion from a single RGB-D image.
In Proceedings of ICPR 2020 - 25th International Conference on Pattern Recognition.
IEEE.
.
(doi:10.1109/ICPR48806.2021.9413252).
Record type:
Conference or Workshop Item
(Paper)
Abstract
Semantic scene completion is the task of predicting a complete 3D representation of volumetric occupancy with corresponding semantic labels for a scene from a single point of view. In this paper, we present EdgeNet, a new end-to-end neural network architecture that fuses information from depth and RGB, explicitly representing RGB edges in 3D space. Previous works on this task used either depth-only or depth with colour by projecting 2D semantic labels generated by a 2D segmentation network into the 3D volume, requiring a two step training process. Our EdgeNet representation encodes colour information in 3D space using edge detection and flipped truncated signed distance, which improves semantic completion scores especially in hard to detect classes. We achieved state-of-the-art scores on both synthetic and real datasets with a simpler and a more computationally efficient training pipeline than competing approaches.
This record has no associated files available for download.
More information
Accepted/In Press date: 22 June 2020
Published date: 15 January 2021
Additional Information:
Funding Information:
This work was supported in part by EPSRC Platform Grant in Visual AI Research EP/P0022529. Dr. de Campos would like to thank FAPDF (fap.df.gov.br) and CNPq grant PQ 314154/2018-3 (cnpq.br) for the financial support to this work. Mr. Dourado would like to thank to TCU (tcu.gov.br) for supporting his PhD studies.
Publisher Copyright:
© 2020 IEEE
Venue - Dates:
25th International Conference on Pattern Recognition, ICPR 2020, , Virtual, Milan, Italy, 2021-01-10 - 2021-01-15
Identifiers
Local EPrints ID: 445075
URI: http://eprints.soton.ac.uk/id/eprint/445075
ISSN: 1051-4651
PURE UUID: 5a7e10d6-8b98-458c-8501-1c2868d8a022
Catalogue record
Date deposited: 19 Nov 2020 17:30
Last modified: 18 Mar 2024 03:56
Export record
Altmetrics
Contributors
Author:
Aloisio Dourado
Author:
Teofilo E. de Campos
Author:
Hansung Kim
Author:
Adrian Hilton
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics