Deep Visual Semantic Embedding with Text Data Augmentation and Word Embedding Initialization

Figure 1

Proposed dual-normalized visual semantic embedding learning framework.