weihancug / SSGF-for-HRRS-scene-classification

experiemntal code for "A semi-supervised generative framework with deep learning features for high-resolution remote sensing image scene classification"

SSGF-for-HRRS-scene-classification

Code to replicate the analyses from the paper A semi-supervised generative framework with deep learning features for high-resolution remote sensing image scene classification

Datasets

The datasets used in the work is UCM, WHU-RS19, NWPU-RESISC45, and AID

1 UCM 2 WHU-RS19 3 NWPU-RESISC45 4 AID

Illustration of code

1. selftraining.py is the implenmtation of selftraining method by using one pre-trained CNN model.

2. co-training.py is the cotraining method, wherein two kind of deep features are utilized.

3. Comparision.py is the comparison methods.

4. plot_confusion_matrix.py is to draw the confusion matrix.