KeypointsDataset

class hybrid_learning.datasets.custom.coco.keypoints_dataset.KeypointsDataset(**kwargs)[source]

Bases: COCODataset

Handler for (a subset of) a COCO keypoints dataset. Input images are the original COCO images. Annotations are the keypoint annotations for the images.

This is essentially a wrapper around a COCO handle following the scheme of BaseDataset and allowing for restriction to specific licenses.

Public Data Attributes:

Inherited from : py: class:COCODataset

`COMMERCIAL_LICENSE_IDS`	IDs of COCO image licenses that allow for commercial use.
`DATASET_ROOT_TEMPL`	Default root directory template for image files that accepts the `split` (`'train'` or `'val'`).
`ANNOTATION_FP_TEMPL`	Default template for the annotation file path that accepts the `split` (`'train'` or `'val'`) and the `root` directory.
`DEFAULT_IMG_SIZE`	Default target size of images to use for the default transforms as `(height, width)`.
`settings`	Return information to init new dataset.
`license_mapping`	The mapping of image IDs to license descriptions and URLs.

Inherited from : py: class:BaseDataset

settings

Return information to init new dataset.

Public Methods:

getitem(i)

Collect the image and the keypoint annotations at position i.

Inherited from : py: class:COCODataset

`subset`(*[, license_ids, body_parts, num, ...])	Restrict the items by the given selection criteria and an optional custom condition.
`shuffle`()	Wrapper around `subset()` that only shuffles the instance.
`to_raw_anns`([description, save_as])	Create the content of a new valid annotations file restricted to the current image IDs.
`copy_to`([root_root, description, overwrite, ...])	Create a new dataset by copying used images and annotations to new root folder.
`load_orig_image`(i)	Load unmodified image by index in dataset.
`descriptor`(i)	Return the image file name for the item at index `i`.
`image_filepath`(i)	Path to image file at index `i`.
`image_attribution`(i)	Get attribution information for image at index `i`.
`image_meta`(i)	Load the dict with meta information for image at index `i`.
`raw_anns`(i)	Return the list of raw annotations for image at index `i`.
`getitem`(i)	Collect the image and the keypoint annotations at position `i`.

Inherited from : py: class:BaseDataset

`getitem`(i)	Collect the image and the keypoint annotations at position `i`.
`descriptor`(i)	Return the image file name for the item at index `i`.

Special Methods:

__init__(**kwargs)

Init.

Inherited from : py: class:COCODataset

`__init__`(**kwargs)	Init.
`__len__`()	Length is given by the length of the index mapping.

Inherited from : py: class:BaseDataset

`__init__`(**kwargs)	Init.
`__len__`()	Length is given by the length of the index mapping.
`__getitem__`(idx)	Get item from `idx` in dataset with transformations applied.
`__repr__`()	Nice printing function.

Inherited from : py: class:Dataset

`__getitem__`(idx)	Get item from `idx` in dataset with transformations applied.
`__add__`(other)

__init__(**kwargs)[source]

Init.

Parameters

spec – Specification
kwargs – Arguments for super class

classmethod get_default_transforms(img_size, device=None)[source]

Return the default transformation, which is pad and resize.

Parameters

img_size (Tuple[int, int]) –
device (Optional[device]) –

Return type

TupleTransforms

getitem(i)[source]

Collect the image and the keypoint annotations at position i. Used for __getitem__().

Parameters: i (int) –
Return type: Tuple[Image, List[Dict]]

__parameters__ = ()

after_cache_transforms: Callable: Transformation function applied after consulting the cache (no matter, whether the tuples was retrieved from cache or not). Use these transformations instead of transforms to ensure the transformation is always applied, regardless of caching. By default, tensor gradients are disabled and tensors are moved to a common device (see _get_default_after_cache_trafo()).

coco: COCO: Internal COCO handle.

dataset_root: str: Assuming the dataset is saved in some storage location, a root from which to navigate to the dataset information.

img_ann_ids: List[Tuple[int, List[int]]]: Mapping of indices in this dataset to COCO image and annotation IDs. Each entry in the list is a tuple of the form (image_id, [annotation_id, ...]) where the annotations belong to the corresponding image.

split: Optional[DatasetSplit]: Optional specification what use-case this dataset is meant to represent, e.g. training, validation, or testing.

transforms: Callable: Transformation function applied to each item tuple before return. Applied in __getitem__(). Default transformations are sub-class-specific. Items transformed using transforms can be cached by setting transforms_cache. If the transformations should be applied always, regardless of caching, use after_cache_transforms.

transforms_cache: Optional[Cache]: Cache for the transformed (input, target) tuples. If set, __getitem__() will first try to load the tuple from cache before loading and transforming it normally. Items not in the cache are put in there after transforms is applied.