Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames