Deployable 3D Semantic Segmentation for AI-Assisted Bridge Inspection: Benchmarking Modern Point-Cloud Models Under an 8 GB GPU Constraint
Infrastructures · 17.11.2026 · 11 (11) · 425
DOI: 10.3390/infrastructures11110425 ↗Kurz gefasst
Fachartikel in der Zeitschrift Infrastructures, online erschienen am 17.11.2026. Verfasst von Kenta Itakura und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Japan Society for the Promotion of Science, Council for Science, Technology and Innovation, Japan Construction Information Center.
Abstract (Originalsprache)
AI-powered damage identification and condition assessment of civil infrastructure increasingly rely on accurate semantic understanding of three-dimensional sensing data. In bridge inspection and post-disaster assessment, component-level semantic segmentation of point clouds provides essential structural context for localizing and interpreting potential damage. However, modern 3D segmentation models are often evaluated mainly by accuracy under high-end computing environments, whereas field-oriented damage assessment may require local processing with limited computational resources. This paper studies deployable 3D semantic segmentation for real-world bridge point clouds under an explicit 8 GB dedicated-VRAM single-GPU constraint. We benchmark representative architectures on five heterogeneous bridge datasets, reporting not only mIoU and overall accuracy but also training time, peak GPU memory, and cross-dataset generalization to a post-disaster bridge dataset. The results show that deployability is not determined by model capacity, memory footprint, or single-dataset accuracy alone; instead, practical deployment for AI-assisted infrastructure inspection and damage-assessment workflows requires a balance of accuracy, runtime, memory use, and robustness to domain variation.
Förderung
- Japan Society for the Promotion of Science
- Council for Science, Technology and Innovation
- Japan Construction Information Center
Zitieren
APA
Itakura, K., Orimoto, S., Murai, T., & Chun, P. (2026). Deployable 3D Semantic Segmentation for AI-Assisted Bridge Inspection: Benchmarking Modern Point-Cloud Models Under an 8 GB GPU Constraint. Infrastructures, 11(11), 425. https://doi.org/10.3390/infrastructures11110425
Vancouver
Itakura K, Orimoto S, Murai T, Chun P. Deployable 3D Semantic Segmentation for AI-Assisted Bridge Inspection: Benchmarking Modern Point-Cloud Models Under an 8 GB GPU Constraint. Infrastructures. 2026;11(11):425. doi:10.3390/infrastructures11110425
BibTeX
@article{Itakura2026,
author = {Itakura, Kenta and Orimoto, Seiya and Murai, Takumi and Chun, Pang-jo},
title = {Deployable 3D Semantic Segmentation for AI-Assisted Bridge Inspection: Benchmarking Modern Point-Cloud Models Under an 8 GB GPU Constraint},
journal = {Infrastructures},
year = {2026},
volume = {11},
number = {11},
pages = {425},
doi = {10.3390/infrastructures11110425}
}
EUVIMED