Semi-Supervised Cross-Modal Retrieval Based on Discriminative Comapping

Figure 4

The mAP performance for each class on Wikipedia datasets and Pascal Sentence dataset. (a) I2T on wiki-R. (b) T2I on wiki-R. (c) Average on wiki-R. (d) I2T on wiki-W. (e) T2I on wiki-W. (f) Average on wiki-W. (g) I2T on Pascal. (h) T2I on Pascal. (i) Average on Pascal.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)