A novel remote image caption method that use target mask and semantic information, the related datasets and code will be opensourced at here.
First, download the json form [data][(https://drive.google.com/drive/folders/1ZYPBMrXowpShHMItXfLG9ggdD-0ZfVut?usp=drive_link. and mask form [data][ https://drive.google.com/drive/folders/1AK-E9BemytL24tixI7t2y2W4NW5n28Qb?usp=drive_link )], and refer to the [RSICC][(https://github.com/Chen-Yang-Liu/RSICC)] to generate the required data in the data\UCM_images1.
Then python train_SSCPNet_NWPU.py
Finally python eval7_NWPU.py