Table 1 Comparison results of retrieval performance on Fashion200k dataset.

Han et al.³⁰

6.3

19.9

38.3

Image only¹¹

3.5

22.7

43.7

Text only¹¹

1.0

12.3

21.8

Concatenation¹¹

11.9 ± 1.0

39.7 ± 1.0

62.6 ± 0.7

Show and Tell³¹

12.3 ± 1.1

40.2 ± 1.7

61.8 ± 0.9

Param Hashing¹⁷

12.2 ± 1.1

40.0 ± 1.1

61.7 ± 0.8

Relationship ¹⁵

13.0 ± 0.6

40.5 ± 0.7

62.4 ± 0.6

FiLM¹⁶

12.9 ± 0.7

39.5 ± 2.1

61.9 ± 1.9

TIRG¹¹

14.1 ± 0.6

42.5 ± 0.7

63.8 ± 0.8

Zhang et al.¹⁸

17.3 ± 0.6

45.2 ± 0.9

65.7 ± 0.8

Ours

17.7 ± 0.6

46.8 ± 0.6

66.2 ± 0.9