- Article
This paper presents a non-learning-based, seed-dependent, semi-automatic pedestrian candidate generation pipeline for LiDAR point clouds. The proposed method is designed to support 3D annotation workflows by reducing irrelevant candidate clusters whi...