Table 2 Averaged categorization accuracies on the compared models on the aforementioned datasets (%).

From: Scene categorization by Hessian-regularized active perceptual feature selection

Dataset

FWK

FTK

MRH

PM

LLC-SP

SC-SP

OB-SP

SV

SSC

Scene-15

72.1

75.4

67.2

77.6

81.3

82.1

77.1

82.1

87.4

Scene-67

41.6

41.8

34.2

44.5

48.5

47.7

48.6

47.3

51.3

ZJU Aerial

66.8

68.3

62.5

73.3

78.4

78.1

78.1

78.3

82.6

ILSVRC-2010

32.1

30.7

27.4

32.4

38.4

36.3

37.2

37.2

38.4

SUN397

15.3

15.6

14.2

22.3

39.3

39.5

38.0

35.5

40.2

Places205

22.1

22.2

20.6

27.5

31.2

32.3

31.6

31.3

32.2

MSEI

47.5

48.2

50.6

47.3

51.1

54.1

47.5

51.3

52.7

Dataset

IN-CNN

R-CNN

M-CNN

DM-CNN

SPP-CNN

SP-S

SP-GBV

SP-LDA

Mesnil

Scene-15

83.1

87.4

87.3

89.3

92.3

90.5

86.2

87.1

86.4

Scene-67

57.2

68.1

72.3

68.4

65.3

76.2

71.5

72.1

71.8

ZJU Aerial

75.2

79.1

78.2

81.0

78.2

81.2

80.3

81.1

80.6

ILSVRC-2010

35.7

38.4

40.4

40.6

41.3

41.4

40.4

40.5

40.5

SUN397

48.1

47.2

51.2

48.7

52.1

51.7

50.5

51.0

50.5

Places205

40.7

43.7

44.8

45.9

48.3

49.9

48.4

48.1

49.4

MSEI

52.4

50.5

51.4

53.5

55.7

52.6

58.1

61.3

62.1

Dataset

Xiao

Cong

Fast R-CNN

Faster R-CNN

Ours (MKL)

Ours (Softmax)

   

Scene-15

82.8

86.6

90.2

91.2

93.4

92.1

   

Scene-67

71.3

72.1

71.5

74.7

75.6

72.9

   

ZJU Aerial

81.1

80.1

78.6

81.2

84.3

82.6

   

ILSVRC-2010

40.5

41.1

40.8

41.1

44.2

42.7

   

SUN397

50.4

51.2

52.2

52.0

56.3

53.2

   

Places205

49.3

48.2

48.3

49.3

52.1

50.1

   

MSEI

59.7

61.5

62.5

64.7

72.4

71.6

   
Back to article page