Here is a list of datasets we've used and there structure. Pictures are "open-source", under various licences. Reach out if you need specifics
Dataset name | link | description |
---|---|---|
27.02.FloodingData.zip | GDrive | 486 pictures of "unaltered" houses, 539 pictures involving floods (+ Venice) of which we consider 159 to be of good quality to train a GAN |
10.03CroppedData.zip | GDrive | Same pictures as 27.02.FloodingData but houses sorted according to presence of Grass or Street |
11.03AllHousesAugmented.zip | GDrive | 3,562 pictures of houses, augmented based on all of the images of unaltered houses from 27.02.FloodingData by reducing dimensionality, cropping and flipping horizontally |
11.03AllHousesFlooded.zip | GDrive | 3,539 pictures of flooding, augmented based on all of the flooding images from 27.02.FloodingData by reducing dimensionality, cropping and flipping horizontally |
11.03BestFloodAugmented | GDrive | 1,175 pictures of flooding, augmented based on the best flooding images from 27.02.FloodingData by reducing dimensionality, cropping and flipping horizontally |