PatchDatasetABC¶
tiatoolbox.models.dataset.abc.PatchDatasetABC
- class PatchDatasetABC[source]¶
Defines abstract base class for patch dataset.
- return_labels¶
__getitem__ will return both the img and its label. If labels is None, None is returned
- Type
bool, False
- preproc_func¶
Preprocessing function used to transform the input data. If supplied, then torch.Compose will be used on the input preprocs. preprocs is a list of torchvision transforms for preprocessing the image. The transforms will be applied in the order that they are given in the list. For more information, use the following link: https://pytorch.org/vision/stable/transforms.html.
Methods
Load an image from a provided path.
Define the pre-processing of this class of loader.
Attributes
Return the current pre-processing function of this instance.
- static load_img(path)[source]¶
Load an image from a provided path.
- Parameters
path (str) – Path to an image file.
- property preproc_func¶
Return the current pre-processing function of this instance.
The returned function is expected to behave as follows: >>> transformed_img = func(img)