Japan landslide dataset for semantic segmentation

This database contains images used for the semantic segmentation of landslide scars from a fully convolutional neural network U-Net.

1. Training dataset: it contains 125 GeoTIFF 8 bits images and associated PNG masks (scars indicated in white and background in black color).

2. Validation dataset: it contains 10 GeoTIFF 8 bits images and associated PNG masks used for U-Net validation step.

3. Test dataset: it contains 10 GeoTIFF 8 bits images and associated PNG masks for testing.

Also, the "SHAPEFILES_LANDSLIDES.rar" file contains the vector layers of the masked images in .shp format.