img = np.concatenate((img,b,b), axis=2)

danfenghong / IEEE_TPAMI_SpectralGPT

Hong, D., Zhang, B., Li, X., Li, Y., Li, C., Yao, J., Yokoya, N., Li, H., Ghamisi, P., Jia, X., Plaza, A. and Gamba, P., Benediktsson, J., Chanussot, J. (2024). SpectralGPT: Spectral remote sensing foundation model. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024. DOI:10.1109/TPAMI.2024.3362475.

161 stars 17 forks source link

img = np.concatenate((img,b,b), axis=2) #13

Closed andremakar closed 4 months ago

andremakar commented 4 months ago

Please tell me, why does the TUM_128 file combine HIS with two "medium" b channels, and not with one? these are two identical layers. the file TUM_128.py , line 96

moonboy12138 commented 4 months ago

This implementation is simply adapted to use the same channel as the pretrained data. You can delete it and use the original 10 channels, which won't affect the model's performance.