Table 3 Comparison of single rubbing retrieval methods on Top-1 accuracy, Top-10 accuracy, MR, and MRR

From: An open benchmark for oracle bone rubbing image retrieval

Category

Method

Top-1

Top-10

MR

MRR

End-to-end

MeCoq 12

3.13%

10.23%

9.42

0.1389

 

HiHPq 43

2.47%

11.38%

9.37

0.1376

 

Dino 44

35.47%

53.13%

5.86

0.4522

Feature-based

SIFT45

81.35%

85.97%

2.35

0.8488

 

SuperGlue25

84.48%

89.27%

2.05

0.8760

 

SGMnet46

79.70%

86.46%

2.41

0.8348

 

LightGlue47

78.38%

89.60%

2.22

0.8354

 

DeDoDe26

79.37%

83.99%

2.54

0.8278

Content-based

MSHRR (ours)

84.81%

90.09%

1.96

0.8817

  1. The bold values indicate the best-performing methods for each evaluation metric, highlighting the highest scores across all compared approaches.
Back to article page