Table 6 The experimental results of the mapping edge feature accuracy between the source and predicted target images.

From: A construction of heterogeneous transfer learning model based on associative fusion of image feature data

Differences in the presence of edge features in the source and predicted target images.

Input Image

Evaluating Indicator

Notebook

Computer

Bicycle

Zebra

Flower

\(K(W)\)

0.02

0.01

0.02

0.01

\(K(V)\)

0.03

0.02

0.02

0.02

\(K(B)\)

0.04

0.03

0.03

0.04

Whether the dimension of public feature Mapping space is maximal

Yes

Yes

Yes

Yes

Accuracy of image edge feature(%)

95.31

95.43

95.76

96.08

Accuracy of image edge feature space (%)

94.87

95.01

94.33

94.75

Accuracy of image edge feature label (%)

93.27

94.11

93.82

94.01

Differences in the presence of edge features in the source and predicted target images.

Input Image

Evaluating Indicator

Notebook

Computer

Bicycle

Zebra

Flower

\(K(W)\)

0.03

0.02

0.03

0.02

\(K(V)\)

0.02

0.02

0.03

0.02

\(K(B)\)

0.04

0.03

0.03

0.04

Whether the dimension of public feature mapping space is maximal

yes

yes

yes

yes

Accuracy of image edge feature(%)

93.44

94.17

93.82

93.55

Accuracy of image edge feature space(%)

91.79

92.22

91.37

93.05

Accuracy of image edge feature label(%)

90.51

90.39

91.17

91.80