VizObj2Vec: Contextual Representation Learning for Visual Objects in Video-frames | IEEE Conference Publication | IEEE Xplore