PatchDataset¶

tiatoolbox . models . dataset . dataset_abc . PatchDataset

class PatchDataset(inputs, labels=None, patch_input_shape=None)[source]¶

Define PatchDataset for torch inference.

Define a simple patch dataset, which inherits from the: torch.utils.data.Dataset class.

Parameters:

inputs (list[str | Path | ndarray] | ndarray)
labels (list[int] | ndarray)
patch_input_shape (IntPair | None)

inputs¶

Either a list of patches, where each patch is a ndarray or a list of valid path with its extension be (“.jpg”, “.jpeg”, “.tif”, “.tiff”, “.png”) pointing to an image.

Type:: list or np.ndarray

labels¶

List of labels for sample at the same index in inputs. Default is None.

Type:: list

patch_input_shape¶

Size of patches input to the model. Patches are at requested read resolution, not with respect to level 0, and must be positive.

Type:: tuple

Examples

>>> # A user defined preproc func and expected behavior
>>> preproc_func = lambda img: img/2  # reduce intensity by half
>>> transformed_img = preproc_func(img)
>>> # create a dataset to get patches preprocessed by the above function
>>> ds = PatchDataset(
...     inputs=['/A/B/C/img1.png', '/A/B/C/img2.png'],
...     labels=["labels1", "labels2"],
...     patch_input_shape=(224, 224),
... )

Initialize PatchDataset.

Methods

Attributes

`inputs`
`labels`