Summary
Indoor mobile laser scanners collect 3D data point by point, but most current datasets label whole scenes or are outdoor focused, leaving single-frame indoor data unlabeled. The paper introduces Semantic-ITC, a large dataset with detailed labels for individual frames collected inside buildings like classrooms and offices. The labeling combined computer vision models with building plans and human editing for accuracy, and it keeps the natural sparse pattern of laser scans. This dataset helps measure how well algorithms can recognize different indoor objects from single laser scan frames, highlighting challenges where object edges and similar objects cause mistakes.
What this means in practice
- •For robotics teams: Test and improve robots’ ability to identify rooms and objects from single indoor laser scan frames using this labeled dataset.
- •For game developers: Use accurate labeled indoor LiDAR frames to create realistic and semantically rich indoor 3D environments for virtual reality experiences.
Authors
Haiyang Wu, Muhammad Affan, George Vosselman, Ville Lehtola
Abstract
Semantic labels for indoor mobile laser scanning (MLS) frames remain largely absent from current point cloud semantic segmentation benchmarks, which mainly focus on reconstructed indoor scenes or outdoor LiDAR perception. This paper introduces Semantic-ITC, to the best of our knowledge the first public dataset and benchmark for frame-wise indoor MLS semantic segmentation. The dataset contains 52 indoor sequences, 79,108 MLS frames, and 1.23 billion labeled points collected in classrooms, corridors, meeting rooms, offices, and study areas. Labels are attached directly to measured LiDAR points in each frame using 16 semantic classes covering structural elements, furniture, room equipment, vegetation, and other indoor objects. Semantic-ITC preserves the sparse, non-uniform, and frame-wise sampling pattern of indoor MLS, making it distinct from scene-level reconstructed point clouds and mesh-based indoor datasets. The annotations are produced by a hybrid workflow that combines predictions from a visual foundation model applied to synchronized RGB images, structural information from BIM, and manual refinement, with the final labels assigned to the original LiDAR frames. A single-frame benchmark is provided, and the best baseline reaches 79.27\% mIoU. Remaining errors are concentrated around object boundaries and ambiguous indoor classes, indicating the challenges of indoor MLS segmentation under sparse frame geometry and long-tailed class distributions. The dataset provides a public benchmark for evaluating semantic segmentation directly on measured indoor MLS frames and supports future studies on frame-wise indoor MLS semantic segmentation.