Table 2.
This table shows the statistical significance of the differences among the two proposed classifiers by a different number of training images.
No of images | 4 | 6 | 8 | 10 | 12 | 15 | 20 | 30 | 50 | 100 |
---|---|---|---|---|---|---|---|---|---|---|
p value | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 |
χ2 | 0963.26 | 1888.42 | 2101.88 | 2513.66 | 1044.49 | 1590.29 | 1552.43 | 1056.37 | 906.48 | 959.85 |
No of images | 200 | 500 | 1k | 2k | ||||||
p value | 0.00 | 0.00 | 0.38 | 1.00 | ||||||
χ2 | 280.09 | 209.58 | 157.85 | 102.98 |
It can be appreciated that the differences are significant (pvalue < 0.01) for training image numbers lower than 500 whereas classifier differences are not significant with larger number of training images.