using objects detected or recognised in the video content