Skip to main content
Account

Table 5 Main module ablation study

From: AMCA-SGG: Adaptive multi-prototype learning and cross-modal alignment for unbiased scene graph generation

Modules

SGCLS

MPC

HRRM

VTA

mR@20

mR@50

mR@100

   

19.5

22.5

23.5

\(\checkmark \)

  

21.5

23.7

24.5

\(\checkmark \)

\(\checkmark \)

 

23.3

26.7

27.4

\(\checkmark \)

 

\(\checkmark \)

22.6

25.3

26.0

\(\checkmark \)

\(\checkmark \)

\(\checkmark \)

26.0

28.5

29.2

  1. The best results are highlighted in bold