Text this: Video understanding using multimodal deep learning