The KIT IIIT Motion Capture Head Pose Dataset consists von videos and corresponding ground truth data for the head pose of individuals shown in the video sequences. The video data is recorded in RGB24 format with a resolution of 800x600 pixels. Calibration data in form of the intrinsic camera parameters are provided.
Below, a sample video of the dataset is shown. The displayed video differs from the videos in the dataset as it has an decresed resolution of 640x480 pixels and is compressed utilizing the mpeg4 codec for representation issues.