Learning Deep Representations of Fine-Grained Visual Descriptions | IEEE Conference Publication | IEEE Xplore