Coupling Self-Distillation with Test Time Augmentation for effective LiDAR-Based 3D Semantic Segmentation
No Thumbnail Available
Date
2025
Journal Title
Journal ISSN
Volume Title
Publisher
The Eurographics Association
Abstract
Effective 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-main
Description
@inproceedings{10.2312:3dor.20251201,
booktitle = {Eurographics Workshop on 3D Object Retrieval},
editor = {Guerrero, Paul and Pratikakis, Ioannis and Veltkamp, Remco},
title = {{Coupling Self-Distillation with Test Time Augmentation for effective LiDAR-Based 3D Semantic Segmentation}},
author = {Antonarakos, Dimitrios and Zamanakos, Georgios and Papadeas, Ilias and Pratikakis, Ioannis},
year = {2025},
publisher = {The Eurographics Association},
ISSN = {1997-0471},
ISBN = {978-3-03868-280-6},
DOI = {10.2312/3dor.20251201}
}