utils.datasets.STPropheseeDataset

class STPropheseeDataset(gt_files: List[str], data_files: List[str], time_step: int, num_steps: int, num_load_file: int, name: str)

Bases: PropheseeDatasetBase

Single-target Prophesee gen1 and 1mpx iterable datasets

Parameters:

• gt_files (List[str]) – List of ground truth file paths.

• data_files (List[str]) – List of data file paths.

• time_step (int) – Time between frames.

• num_load_file (int) – Number of examples loaded at a time.

• name (str) – The name of the dataset to download. Supported gen1 and 1mpx.

Raises:

ValueError – Invalid dataset name.

Methods

parse_data	Transforms events into a video stream
samples_generator	Creates a new sample generator

Attributes

box_size_threshold	Minimum acceptable box size relative to frame area
events_threshold	Minimum average number of events in a sample to use it

box_size_threshold: float = 0.01

Minimum acceptable box size relative to frame area

events_threshold: int = 4000

Minimum average number of events in a sample to use it

parse_data(gt_boxes: Tensor, events_loader: PSEELoader) → Tuple[Tuple[Tensor, Tensor] | None, bool]

Transforms events into a video stream

samples_generator() → Generator[Tuple[Tensor, Tensor], None, None]

Creates a new sample generator