Abdielfer

Dataset split.

1

I notice percent in the split dataset do not match the expected proportions. It seems split is made before filtering patches by min_annot_perc.

To many patches after tiling (for flood modelling interest)

1

After tiling, the destination folder contains all possible patches from the raster image, event when there is not ground truth (GT) labels matching (< min_annot_perc). Flood happens close to water...

Tiling: (re)implement overlap between patches

1

A patch stride is present in our Tiler class attribute, but it not yet implemented: https://github.com/NRCan/geo-deep-learning/blob/develop/tiling_segmentation.py#L71 To complete the implementation, it should be used to set the stride for the...