vp_suite.datasets.human36m
- class Human36MDataset(split, **dataset_kwargs)
Bases:
vp_suite.base.base_dataset.VPDatasetDataset class for the Videos of the dataset “Human 3.6M”, as encountered in “Human3.6M: Large Scale Datasets and Predictive Methods for 3D Human Sensing in Natural Environments” by Ionescu et al. (http://vision.imar.ro/human3.6m/pami-h36m.pdf).
Each sequence depicts a human actor in a room equipped with different cameras and sensors. The actor is one of several different scenarios such as “Discussion”, “Sitting” or “Smoking”.
- ACTION_SIZE: int = 0
The size of the action vector per frame (If the dataset provides no actions, this value is 0).
- ALL_SCENARIOS = ['Directions', 'Discussion', 'Eating', 'Greeting', 'Phoning', 'Photo', 'Posing', 'Purchases', 'Sitting', 'SittingDown', 'Smoking', 'TakingPhoto', 'Waiting', 'WalkDog', 'WalkTogether', 'Walking', 'WalkingDog']
All recorded scenarios of the dataset.
- DATASET_FRAME_SHAPE: (<class 'int'>, <class 'int'>, <class 'int'>) = (1000, 1000, 3)
For Human 3.6M, some sequences come in a shape of (1002, 1000, 3). They’re resized during loading.
- DEFAULT_DATA_DIR: pathlib.Path = PosixPath('/home/runner/work/vp-suite/vp-suite/vp-suite-data/data/human36m')
The default save location of the dataset files.
- FPS = 50
Frames per Second.
- IS_DOWNLOADABLE: str = 'With Registered Account'
A string identifying whether the dataset can be (freely) downloaded.
- REFERENCE: str = 'http://vision.imar.ro/human3.6m/description.php'
The reference (publication) where the original dataset is introduced.
- SKIP_FIRST_N = 25
Some of the sequences start with a bit of idling from the actor. Therefore, the first few frames of each sequence are discarded.
- VALID_SPLITS = ['train', 'val', 'test']
The valid arguments for specifying splits.
- __init__(split, **dataset_kwargs)
Initializes the dataset loader by determining its split and extracting and processing all dataset attributes from the parameters given in dataset_kwargs.
- Parameters
split (str) – The dataset’s split identifier (i.e. whether it’s a training/validation/test dataset)
**dataset_kwargs (Any) – Optional dataset arguments for image transformation, value_range, splitting etc.
- classmethod download_and_prepare_dataset()
Downloads the specific dataset, prepares it for the video prediction task (if needed) and stores it in a default location in the ‘data/’ folder. Implemented by the derived dataset classes.
- scenarios = None
Scenarios chosen for the current dataset instance (defaults to self.ALL_SCENARIOS)