From: Classification of hyperspectral imagery with neural networks: comparison to conventional tools
194-band | 30-band | 13-band | ||||
---|---|---|---|---|---|---|
User | Producer | User | Producer | User | Producer | |
Class | Accuracy | Accuracy | Accuracy | |||
A | 93.31 | 90.00 | 97.75 | 98.06 | ||
B | 75.00 | 60.00 | 100.00 | 18.50 | ||
C | 58.49 | 46.50 | 68.36 | 87.50 | ||
D | 75.74 | 84.76 | 83.47 | 93.81 | ||
E | 99.44 | 84.76 | 39.11 | 100.00 | ||
F | 35.87 | 82.50 | 10.03 | 100.00 | ||
G | 100.00 | 50.29 | 36.75 | 24.86 | ||
H | 72.91 | 83.18 | 78.80 | 77.73 | ||
I | 63.50 | 87.00 | 23.54 | 93.00 | ||
J | 97.50 | 93.60 | 100.00 | 0.40 | ||
K | 80.81 | 64.00 | 93.94 | 49.60 | ||
L | 74.09 | 85.51 | 71.08 | 95.33 | ||
M | 40.38 | 10.50 | 00.00 | 00.00 | ||
N | 85.25 | 52.00 | 100.00 | 5.00 | ||
O | 57.49 | 60.65 | 48.41 | 59.03 | ||
P | 82.43 | 81.33 | 78.19 | 83.67 | ||
Q | 87.76 | 94.55 | 100.00 | 11.82 | ||
R | 82.46 | 94.00 | 100.00 | 24.00 | ||
S | 89.06 | 95.00 | 100.00 | 33.33 | ||
T | 89.71 | 93.85 | 94.83 | 84.62 | ||
U | 48.48 | 95.50 | 81.60 | 66.50 | ||
V | 25.53 | 72.00 | 100.00 | 12.00 | ||
W | 51.53 | 59.00 | 84.92 | 76.00 | ||
Accuracy (%) = N/A | Accuracy (%) = 72.53 | Accuracy (%) = 59.34 | ||||
κ = N/A | κ = 0.7108 | κ = 0.5726 |