liom_toolkit.segmentation.vseg.make_dataset module

liom_toolkit.segmentation.vseg.make_dataset.make_train_val(train_images: list, val_images: list, norm: bool, stride: int, width: int, output_dir: str, threshold: int) → None

Make training and validation datasets

Parameters:

• train_images (list) – The training images

• val_images (list) – The validation images

• norm (bool) – Normalize the images

• stride (int) – The stride of the patches

• width (int) – The width of the patches

• output_dir (str) – The output directory

• threshold (int) – The threshold for removing background tiles

Returns:

None