Semi-Supervised Cross-Modal Retrieval Based on Discriminative Comapping

Figure 1

Using an image to retrieve texts and a text to retrieve images.