Skip to main content

Table 1 Performance (%) comparisons of different methods on EYB in size of 8 × 8 pixels

From: Class-specific kernel linear regression classification for face recognition under low-resolution and illumination variation conditions

Methods\testing faces

Subset 2

Subset 3

Subset 4

Subset 5

Average

PCA+Euclidean

70.18

19.52

2.63

2.63

23.74

PCA+Mahalanobis

66.67

34.17

19.29

18.42

34.64

KPCA-p

75.57

22.85

2.13

2.77

25.83

KPCA-g

76.67

25.48

3.57

2.91

27.16

LDA+Euclidean

100.00

89.91

26.50

4.02

55.11

LDA+Mahalanobis

88.33

72.50

27.86

7.89

49.15

KLDA-p

98.68

64.69

24.66

4.57

48.15

KLDA-g

99.56

74.34

26.54

4.71

51.29

LRC

100.00

98.03

51.69

11.77

65.37

RLRC

100.00

98.68

60.34

22.16

70.30

SRC

95.83

41.67

17.86

23.68

44.76

LPP

100.00

80.00

30.71

13.68

50.10

NPE

97.50

49.17

19.29

15.79

45.44

IPCR

100.00

96.67

56.43

16.84

67.49

URC

100.00

97.50

57.14

21.05

68.92

LDRC

100.00

75.83

39.29

17.37

58.12

LBP

96.67

70.00

31.43

23.68

55.40

KLRC-p

100.00

98.68

77.82

46.26

80.69

KLRC-g

100.00

98.68

79.51

45.01

80.80

  1. Italicized data means the proposed methods' results