SSDGL: A Spectral-Spatial-Dependent Global Learning Framework for Insufficient and Imbalanced Hyperspectral Image Classification (TCYB2021) https://ieeexplore.ieee.org/document/9440852
GNU General Public License v3.0
52
stars
15
forks
source link
Question about the ablation study of GCL and GJAM module #2
Thank you very much for the code you shared. I have a question about the ablation study.
We carried experiments on 5% training samples of Indian Pines dataset to validate the effectiveness of GCL and GJAM module. Experiments show that under the same HB-WL strategy of this paper, the classification accuracy containing GCL and GJAM module (OA:99.51%, AA:99.81%) is less than that without GCL and GJAM module (OA:99.69%, AA:99.85%). This suggest GCL and GJAM module may be unimportant. Why is this?
Thank you very much for the code you shared. I have a question about the ablation study. We carried experiments on 5% training samples of Indian Pines dataset to validate the effectiveness of GCL and GJAM module. Experiments show that under the same HB-WL strategy of this paper, the classification accuracy containing GCL and GJAM module (OA:99.51%, AA:99.81%) is less than that without GCL and GJAM module (OA:99.69%, AA:99.85%). This suggest GCL and GJAM module may be unimportant. Why is this?