Table 1 Data augmentation performance evaluation on Handwritten Medical Term Corpus.

From: An online cursive handwritten medical words recognition system for busy doctors in developing countries for ensuring efficient healthcare service delivery

SL.

Method

Model accuracy (%)

Data size

Avg

Max

Min

1.

No data expansion

73.4

90.8

19.0

15,911

2.

Rotate

83.8

95.6

3.33

1,591,100

3.

Shift

91.3

93.9

87.8

1,591,100

4.

Stretch

91.4

95.7

87.5

1,591,100

5.

Rotate + Shift

89.5

95.4

68.0

1,591,100

6.

Rotate + Stretch

90.0

95.9

77.4

1,591,100

7.

Shift + Stretch

88.8

94.2

78.0

1,591,100

8.

Rotate + Shift + Stretch

93.0

94.5

92.1

1,591,100

  1. The bold values represent the highest accuracy achieved by each data augmentation method.