Research Article
Diagnosis System of Microscopic Hyperspectral Image of Hepatobiliary Tumors Based on Convolutional Neural Network
Table 2
Comparison results of S4CNN with different methods in the Indian Pines data set.
| Method | Type | SVM | 3DCNN | BASE-3DCNN | SSRN | S4CNN |
| Corn-notill | 87.33 | 83.46 | 69.55 | 30.44 | 100.00 | Corn-min | 80.13 | 92.28 | 97.54 | 98.49 | 95.41 | Corn | 52.31 | 88.13 | 87.96 | 97.89 | 97.43 | Grass-pasture | 91.30 | 87.18 | 95.43 | 92.51 | 99.58 | Grass-trees | 99.73 | 91.36 | 97.24 | 96.73 | 98.07 | Grass-pasture-mowed | 67.88 | 35.71 | 100.00 | 100.00 | 100.00 | Hay-windrowed | 100.00 | 100.00 | 100.00 | 100.00 | 100.00 | OA (%) | 88.70 | 93.15 | 96.15 | 97.12 | 97.71 | AA (%) | 81.56 | 83.63 | 94.52 | 89.64 | 96.87 | Kappa × 100 | 87.06 | 92.16 | 95.62 | 96.72 | 97.39 |
|
|