neodroidvision.data.segmentation.clouds.CloudSegmentationDataset

class neodroidvision.data.segmentation.clouds.CloudSegmentationDataset(csv_path: Path, image_data_path: Path, subset: SplitEnum = SplitEnum.training, transp=True, N_FOLDS=10, SEED=246232)[source]

Bases: Dataset

description

__init__(csv_path: Path, image_data_path: Path, subset: SplitEnum = SplitEnum.training, transp=True, N_FOLDS=10, SEED=246232)[source]

Methods

__init__(csv_path, image_data_path[, ...])

fetch_masks(image_name)

Create mask based on df, image name and shape.

no_info_mask(img)

param img

plot_training_sample()

Wrapper for visualize function.

training_augmentations()

Returns:

validation_augmentations()

Add paddings to make image shape divisible by 32

visualise(image, mask[, original_image, ...])

Plot image and masks.

visualise_prediction(processed_image, ...[, ...])

Plot image and masks.

Attributes

categories

image_size

image_size_T

mean

predictor_channels

predictors_shape

predictors_shape_T

response_channels

response_shape

response_shape_T

std

fetch_masks(image_name: str)[source]

Create mask based on df, image name and shape.

static no_info_mask(img)[source]
Parameters

img –

Returns:

plot_training_sample()[source]

Wrapper for visualize function.

training_augmentations()[source]

Returns:

validation_augmentations()[source]

Add paddings to make image shape divisible by 32

static visualise(image, mask, original_image=None, original_mask=None)[source]

Plot image and masks. If two pairs of images and masks are passes, show both.

static visualise_prediction(processed_image, processed_mask, original_image=None, original_mask=None, raw_image=None, raw_mask=None)[source]

Plot image and masks. If two pairs of images and masks are passes, show both.