Skip to main content

Table 2 Detail of experimental parameters of the proposed method

From: BoVW model based on adaptive local and global visual words modeling and log-based relevance feedback for semantic retrieval of the images

Datasets/visual vocabularies

Percentage of the images selected for training

Percentage of the images selected for testing

Corel 1000

70%

30%

Corel 1500

50%

50%

Scene 15

60%

40%

Holidays

60%

40%

Size of the BFGF-HOG visual vocabularies

10, 25, 50, 100, 200, 300, 400, 500, 600 (unit is visual word)

Size of the GSURF visual vocabularies

10, 25, 50, 100, 200, 300, 400, 500, 600 (unit is visual word)

Size of complementary visual vocabulary

20, 50, 100, 200, 400, 600, 800, 1000, 1200 (unit is visual word)