Coupling Self-Distillation with Test Time Augmentation for effective LiDAR-Based 3D Semantic Segmentation

Simple item page
dc.contributor.authorAntonarakos, Dimitriosen_US
dc.contributor.authorZamanakos, Georgiosen_US
dc.contributor.authorPapadeas, Iliasen_US
dc.contributor.authorPratikakis, Ioannisen_US
dc.contributor.editorGuerrero, Paulen_US
dc.contributor.editorPratikakis, Ioannisen_US
dc.contributor.editorVeltkamp, Remcoen_US
dc.date.accessioned2025-08-29T08:12:26Z
dc.date.available2025-08-29T08:12:26Z
dc.date.issued2025
dc.description.abstractEffective 3D perception is fundamental for spatial awareness and safe navigation in modern autonomous systems, with 3D semantic segmentation of LiDAR point clouds being a critical perception task. Recent progress in 2D vision highlights the potential of non-architectural training and inference strategies to further boost model performance. Inspired by consistency-based learning and self-distillation, this work employs such a training pipeline for robust 3D semantic segmentation in street scene understanding. Specifically, we incorporate a teacher-student knowledge self-distillation framework that integrates Test-Time Augmentation to enhance the quality of the soft labels generated by the teacher model during training and to improve inference performance. We present a comparative study on the effectiveness of the employed framework across both convolutional and attention-enhanced networks. Experimental results on the Street3D benchmark dataset demonstrate that the adopted training framework coupled with attention-enhanced networks compares favorably with the state-of-the-art for 3D semantic segmentation in the context of autonomous driving. Code is available at https://github.com/DUTH-VCG/Self_Distillation_with_TTA-mainen_US
dc.description.sectionheadersShort Papers
dc.description.seriesinformationEurographics Workshop on 3D Object Retrieval
dc.identifier.doi10.2312/3dor.20251201
dc.identifier.isbn978-3-03868-280-6
dc.identifier.issn1997-0471
dc.identifier.pages7 pages
dc.identifier.urihttps://doi.org/10.2312/3dor.20251201
dc.identifier.urihttps://diglib.eg.org/handle/10.2312/3dor20251201
dc.publisherThe Eurographics Associationen_US
dc.rightsAttribution 4.0 International License
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/
dc.titleCoupling Self-Distillation with Test Time Augmentation for effective LiDAR-Based 3D Semantic Segmentationen_US
Files
Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
3dor20251201.pdf
Size:
304.16 KB
Format:
Adobe Portable Document Format
Download
Collections
3DOR 2025 - Short Papers