Hyperspectral Semantic Segmentation on HSI Drive v2.0 (test)
97.44Accuracy (mu)HyperVision
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| HyperVisionTrainable Params=15.7 M2026.05 | 97.44 | 89.67 | 90.7 | 83.51 | |
| HyperFreeTrainable Params=15.7 M2026.05 | 96.56 | 87.52 | 88.22 | 79.75 | |
| RU-NetTrainable Params=21.1 M2026.05 | 95.53 | 77.09 | 79.36 | 68.61 | |
| U-NetTrainable Params=17.3 M2026.05 | 94.74 | 75.27 | 76.7 | 65.37 | |
| DeepLabV3+ (ResNet101)Trainable Params=58.8 M2026.05 | 94.23 | 73.91 | 74.98 | 63.46 | |
| DeepLabV3+ (MobileNet)Trainable Params=5.2 M2026.05 | 93.79 | 73.16 | 75.5 | 63.57 | |
| CGRSeg (False-color)Trainable Params=19.1 M2026.05 | 81.02 | 48.19 | 58.06 | 40.96 |