Visual Grounding in Video for Unsupervised Word Translation | IEEE Conference Publication | IEEE Xplore