Skip to main content

Table 8 Results of the comparison in the generation processes

From: First-person reading activity recognition by deep learning with synthetically generated images

 

None

P

D

T

PD

DT

PT

PDT

Avg.

Without R

56.7

41.4

62.0

71.3

60.5

79.5

71.6

81.2

65.5

With R

50.1

43.4

50.6

73.5

64.9

79.2

66.2

79.2

63.4

Differences

−6.6

2.0

−11.4

2.2

4.4

−0.3

−5.4

−2.0

−2.1

 

None

R

D

T

RD

DT

RT

RDT

Avg.

Without P

56.7

50.1

62.0

71.3

50.6

79.5

73.5

79.2

65.4

With P

41.4

43.4

60.5

71.6

64.9

81.2

66.2

79.2

63.6

Differences

−15.3

−6.7

−1.5

0.3

14.3

1.7

−7.3

0.0

−1.8

 

None

R

P

T

RP

PT

RT

RPT

Avg.

Without D

56.7

50.1

41.4

71.3

43.4

71.6

73.5

66.2

59.3

With D

62.0

50.6

60.5

79.5

64.9

81.2

79.2

79.2

69.6

Differences

5.3

0.5

19.1

8.2

21.5

9.6

5.7

13.0

10.4

 

None

R

P

D

RP

PD

RD

RPD

Avg.

Without T

56.7

50.1

41.4

62.0

43.4

60.5

50.6

64.9

53.7

With T

71.3

73.5

71.6

79.5

66.2

81.2

79.2

79.2

75.2

Differences

14.6

23.4

30.2

17.5

22.8

20.7

28.6

14.3

21.5

  1. Differences in each comparison about a process X show the values by substraction of scores with X from ones without X