Scaffolding worker IMU time-series dataset for deep learning-based construction site behavior recognition
Abstract: Highlights•SWIT dataset is developed for unsafe behavior recognition in construction.•Nine common unsafe behaviors in scaffolding were selected using safety guidelines.•The benchmark provides a reference for selecting deep learning-based behavior recognition algorithms.•The limitations of the dataset are discussed, and future improvements are proposed.
External IDs:dblp:journals/aei/ParkSJKCP25
Loading