Source code for "PolSAR Image Classification Using Complex-Valued Multiscale Attention Vision Transformer (CV-MsAtViT)" Accepted for publication in International Journal of Applied Earth Observation and Geoinformation
The paper can be accessed through: https://www.sciencedirect.com/science/article/pii/S1569843225000597
Three PolSAR datasets were utilized to assess the performance of the CV-MsAtViT method in this study. Flevoland, San Francisco, and Oberpfaffenhofen. link to the datasets along their class maps is available at: https://mega.nz/folder/WhgT1L4S#PnMttCUpjtwkD8qTEdwZsw
Python 3.9.18, Tensorflow (and Keras) 2.10.0, cvnn 2.0, Tensorflow Probability 0.18.0
To quantitatively measure the proposed CV-MsAtViT model, three evaluation metrics are employed to verify the effectiveness of the algorithm, Overall Accuracy (OA), Average Accuracy (AA) and Cohen's Kappa (k). Also, Each class accuracy has been reported.
@article{ALKHATIB2025104412, title = {PolSAR image classification using complex-valued multiscale attention vision transformer (CV-MsAtViT)}, journal = {International Journal of Applied Earth Observation and Geoinformation}, volume = {137}, pages = {104412}, year = {2025}, issn = {1569-8432}, doi = {https://doi.org/10.1016/j.jag.2025.104412}, url = {https://www.sciencedirect.com/science/article/pii/S1569843225000597}, author = {Mohammed Q. Alkhatib}, }
Feel free to contact me on: [email protected]