core.services.data.dataset.preprocessing.classification_dataset_preparator

classification_dataset_preparator

Classes:

Name	Description
ClassificationBaseDatasetPreparator	Prepares and organizes dataset images into directories based on their classification categories.

ClassificationBaseDatasetPreparator(dataset, destination_dir)

Prepares and organizes dataset images into directories based on their classification categories.

This class takes a dataset with category and annotation information in COCO format. It organizes the dataset by creating a directory for each category and moves the images into their respective category directories, which is often required for classification tasks in deep learning frameworks.

Attributes:

Name	Type	Description
dataset	BaseDataset	The context of the dataset including paths and COCO file.
destination_dir	str	The target directory where the images will be moved and organized.

Parameters:

Name	Type	Description	Default
dataset	BaseDataset	The context of the dataset to organize.	required
destination_dir	str	The directory where the organized images will be stored.	required

Methods:

Name	Description
organize	Organizes the dataset by creating category directories and moving images.

dataset = dataset instance-attribute

destination_dir = destination_dir instance-attribute

organize()

Organizes the dataset by creating category directories and moving images.

Extracts category information from the COCO file, maps images to their categories, and organizes the images into the respective category directories. Cleans up the original image directory and annotations directory after moving the images.

Returns:

Name	Type	Description
CocoDataset	CocoDataset	The updated dataset with the new image directory.