The dataset is composed of image sequences acquired with 6 fixed cameras that are synchronized and calibrated. It depicts 20 subjects that are walking, in the atrium of a building, and that wear the same uniform. We provide ground-truth trajectories for each subject.
Each video is roughly ~11 minutes long and was taken at ~24fps using a GoPro camera (~16800 images for each view). We provide the intrinsic and extrinsic parameters for each camera and the ground-truth position of each subject on the ground.
contact: [email protected]