Skip to main content

Table 1 Recognition accuracy (%) of various STIP-based approaches in comparison with other approaches on downsampled versions of the KTH dataset

From: Exploiting textures for better action recognition in low-quality videos

Method

SD2

SD3

SD4

TD2

TD3

TD4

STIP (baseline)

86.85

80.37

75.56

88.24

82.31

78.98

STIP+LBP-TOP

85.19

82.04

77.59

88.43

82.41

81.20

STIP+LPQ-TOP

87.41

80.19

76.30

87.41

81.85

79.81

STIP+BSIF-TOP

88.80

85.28

81.67

88.70

86.11

84.54

STIP+Deep object [35]

82.41

82.41

81.48

82.41

80.56

80.09

LTTS+DS model [34]

82.22

83.78

80.00

–

–

–