A CNN-RNN Framework for Image Annotation from Visual Cues and Social Network Metadata | IEEE Conference Publication | IEEE Xplore