Transformer-based Geometric Point Cloud Compression with Local Neighbor Aggregation

Luo, Z; Jia, W; Perry, S

Transformer-based Geometric Point Cloud Compression with Local Neighbor Aggregation

Luo, Z Jia, W Perry, S

Permalink

Publisher:: Institute of Electrical and Electronics Engineers (IEEE)
Publication Type:: Conference Proceeding
Citation:: 2023 International Conference on Digital Image Computing: Techniques and Applications, DICTA 2023, 2023, 00, pp. 223-228
Issue Date:: 2023-01-01

Embargoed

	Filename	Description	Size
	Transformer-based Geometric Point Cloud Compression with Local Neighbor Aggregation.pdf	Accepted version	647.04 kB	Adobe PDF	View/Open

Copyright Clearance Process

Recently Added
In Progress
Embargoed
Open Access

This item is currently unavailable due to the publisher's embargo.

The embargo period expires on 29 Jan 2026

Full metadata record

Field	Value	Language
dc.contributor.author	Luo, Z
dc.contributor.author	Jia, W
dc.contributor.author	Perry, S https://orcid.org/0000-0002-2794-3178
dc.date	2023-11-28
dc.date.accessioned	2024-05-03T08:09:05Z
dc.date.available	2024-05-03T08:09:05Z
dc.date.issued	2023-01-01
dc.identifier.citation	2023 International Conference on Digital Image Computing: Techniques and Applications, DICTA 2023, 2023, 00, pp. 223-228
dc.identifier.isbn	9798350382204
dc.identifier.uri	http://hdl.handle.net/10453/178632
dc.description.abstract	Recently, point cloud processing is becoming popular in AI-driven areas as 3D scanners are developing rapidly. However, this kind of data can have a massive file size, causing significant file storage and transmission difficulties. Compressing point clouds is challenging due to the disordered, sparse, and irregular point cloud structures. Therefore, there is a growing need to develop effective methods to compress point clouds while preserving their information. So far, many methods based on voxel and octree structures have been reported. However, these methods suffer from the information loss issue of local details at early stages, especially during the down-sampling step. In addition, while the global attention mechanism of Transformers has strength in capturing long-range dependency features, it has limitations in capturing local geometry position details. To address these issues, we propose a Transformer-based point cloud geometric compression method with a local neighbor aggregation module to preserve local spatial features during compression. Our method is based on the architecture of the autoencoder, and a Local Neighbor Aggregation module will address the local feature-capturing limitations of the global attention and local spatial data loss in Transformers. Compared with other methods, our method achieves an average of 30.49% and 23.67% bitrate savings in terms of PSNR DI and PSNR D2 respectively with a shorter decoding time.
dc.language	en
dc.publisher	Institute of Electrical and Electronics Engineers (IEEE)
dc.relation.ispartof	2023 International Conference on Digital Image Computing: Techniques and Applications, DICTA 2023
dc.relation.ispartof	2023 International Conference on Digital Image Computing: Techniques and Applications (DICTA)
dc.relation.isbasedon	10.1109/DICTA60407.2023.00038
dc.rights	info:eu-repo/semantics/embargoedAccess
dc.title	Transformer-based Geometric Point Cloud Compression with Local Neighbor Aggregation
dc.type	Conference Proceeding
utslib.citation.volume	00
pubs.organisational-group	University of Technology Sydney
pubs.organisational-group	University of Technology Sydney/Faculty of Engineering and Information Technology
pubs.organisational-group	University of Technology Sydney/Faculty of Engineering and Information Technology/School of Electrical and Data Engineering
pubs.organisational-group	University of Technology Sydney/Strength - VI - Visualisation Institute
utslib.copyright.status	embargoed	*
utslib.copyright.embargo	2026-01-29T00:00:00+1000Z
dc.date.updated	2024-05-03T08:09:02Z
pubs.finish-date	2023-12-01
pubs.publication-status	Published
pubs.start-date	2023-11-28
pubs.volume	00

Abstract:

Recently, point cloud processing is becoming popular in AI-driven areas as 3D scanners are developing rapidly. However, this kind of data can have a massive file size, causing significant file storage and transmission difficulties. Compressing point clouds is challenging due to the disordered, sparse, and irregular point cloud structures. Therefore, there is a growing need to develop effective methods to compress point clouds while preserving their information. So far, many methods based on voxel and octree structures have been reported. However, these methods suffer from the information loss issue of local details at early stages, especially during the down-sampling step. In addition, while the global attention mechanism of Transformers has strength in capturing long-range dependency features, it has limitations in capturing local geometry position details. To address these issues, we propose a Transformer-based point cloud geometric compression method with a local neighbor aggregation module to preserve local spatial features during compression. Our method is based on the architecture of the autoencoder, and a Local Neighbor Aggregation module will address the local feature-capturing limitations of the global attention and local spatial data loss in Transformers. Compared with other methods, our method achieves an average of 30.49% and 23.67% bitrate savings in terms of PSNR DI and PSNR D2 respectively with a shorter decoding time.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/178632