Skip to main content

Table 8 Result obtained from dataset D3 (ResNet-50)

From: Ensemble of deep learning and machine learning approach for classification of handwritten Hindi numerals

Digit class

Precision

Recall

F1 score

Test samples

0

100.00

99.60

99.80

500

1

98.59

98.00

98.29

500

2

94.72

96.80

95.75

500

3

96.95

95.40

96.17

500

4

97.43

98.60

98.01

500

5

97.80

97.60

97.70

500

6

95.10

97.00

96.04

500

7

98.60

98.40

98.50

500

8

98.40

98.20

98.30

500

9

97.55

95.40

96.46

500

Overall recognition accuracy (%)

97.50