Modality-Dependent Cross-Modal Retrieval Based on Graph Regularization

Figure 2

MAP scores for each class on two datasets: (a) I2T on the Wikipedia dataset; (b) T2I on the Wikipedia dataset; (c) average MAP on the Wikipedia dataset; (d) I2T on the Pascal sentence dataset; (e) T2I on the Pascal sentence dataset; (f) average MAP on the Pascal sentence dataset.

(a)

(b)

(c)

(d)

(e)

(f)