search

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.
150 stars 16 forks source link

Hyperspectral image #1

Open Bit0r opened 4 months ago

Bit0r commented 4 months ago

Can the pre-trained weights of SpectralGPT be used for semantic segmentation tasks on hyperspectral images? Can SpectralGPT, which is trained on multispectral images (10+ spectral bands), be applied to downstream tasks on hyperspectral images (100+ spectral bands)?

danfenghong commented 4 months ago

Thank you very much for your interest in our work. Yes, sure. Our SpectralGPT can be used for semantic segmentation on hyperspectral images by using a simple tokenizer to adapt the number of spectral bands.

Bit0r @.***> 于2024年3月3日周日 23:09写道：

Can the pre-trained weights of SpectralGPT be used for semantic segmentation tasks on hyperspectral images? Can SpectralGPT, which is trained on multispectral images (10+ spectral bands), be applied to downstream tasks on hyperspectral images (100+ spectral bands)?

— Reply to this email directly, view it on GitHub https://github.com/danfenghong/IEEE_TPAMI_SpectralGPT/issues/1, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFL2GZUOUYAS5YCTFTAZPITYWM4JZAVCNFSM6AAAAABEEA5OCCVHI2DSMVQWIX3LMV43ASLTON2WKOZSGE3DKNBQGMYDIMA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

andremakar commented 3 months ago

Hello, tell me how to do it if you succeed

Can the pre-trained weights of SpectralGPT be used for semantic segmentation tasks on hyperspectral images? Can SpectralGPT, which is trained on multispectral images (10+ spectral bands), be applied to downstream tasks on hyperspectral images (100+ spectral bands)?