Multi-View In-Cabin Monitoring System for Public Transport Vehicles

A new multi-view in-cabin monitoring dataset has been released for public transport vehicles, featuring synchronized RGB and depth images from four cameras and a rotating LiDAR, totaling 9,136 samples with annotations. It includes a calibration and pseudo-labeling pipeline for generating 3D human pose estimates and oriented 3D bounding boxes, along with benchmark support for models like Lift-Splat-Shoot and BEVFusion. This resource is significant for practitioners as it facilitates the development and evaluation of multi-view perception systems in automated public transport settings.