Skip to main content

Table 1 Proposed algorithm results using EPFL dataset

From: A framework of human detection and action recognition based on uniform segmentation and combination of Euclidean distance and joint entropy-based features selection

Dataset

Classifier

Sequence

ACC (%)

TPR

FPR

Recall (%)

FDR (%)

FNR (%)

AUC

CVLAB campus sequences

M-class SVM

S1:Cv0

100

1.00

0.00

100

0.00

0.00

1.00

S1:Cv1

100

1.00

0.00

100

0.00

0.00

1.00

S2:Cv2

98.70

0.98

0.01

98.00

2.00

1.30

0.99

KNN

S1:Cv0

98.20

0.98

0.01

98.25

1.75

1.80

0.98

S2:Cv1

99.00

0.99

0.01

99.15

0.85

1.00

0.98

S2:Cv2

96.90

0.96

0.03

97.00

3.00

3.10

0.96

EBT

S1:Cv0

99.90

0.99

0.00

100

0.00

0.10

1.00

S2:Cv1

99.70

0.99

0.00

100

0.00

0.30

0.99

S2:Cv2

98.10

0.98

0.01

98.20

1.80

1.90

0.99

CVLAB passageway sequences

M-class SVM

S1:Cv0

100

1.00

0.00

100

0.00

0.00

1.00

S2:Cv1

100

1.00

0.00

100

0.00

0.00

1.00

S2:Cv2

98.70

0.98

0.01

98.00

2.00

1.90

0.99

KNN

S1:Cv0

98.20

0.98

0.01

98.25

1.70

1.80

0.98

S2:Cv1

99.00

0.98

0.01

99.15

0.85

1.00

0.99

S2:Cv2

96.90

0.96

0.03

97.00

3.00

3.10

0.96

EBT

S1:Cv0

100

1.00

0.00

100

0.00

0.00

1.00

S2:Cv1

99.80

0.99

0.00

100

0.00

0.20

0.99

S2:Cv2

98.10

0.98

0.01

98.00

2.00

1.90

0.99