which are AI models that can output videos from images. It generates videos from different angles by generating multiple new view videos (4D image matrix) of objects appearing in the input video.