Research Article
Diagnosis System of Microscopic Hyperspectral Image of Hepatobiliary Tumors Based on Convolutional Neural Network
Table 5
Experimental results of different feature extraction strategies on the Salinas data set.
| ā | Un-pooling | Pooling | OA (%) | Feature size | OA (%) | Feature size |
| h1 | 94.38 | 4704 | 98.24 | 96 | h2 | 94.00 | 4800 | 98.82 | 192 | h3 | 93.70 | 3456 | 97.86 | 384 | h1,h2 | 84.26 | 9504 | 98.99 | 288 | h1,h3 | 94.17 | 8160 | 98.23 | 480 | h2,h3 | 94.02 | 8256 | 98.46 | 576 | h1,h2,h3 | 94.17 | 12960 | 98.60 | 672 |
|
|