vis4d.data.cbgs

Class-balanced Grouping and Sampling for 3D Object Detection.

Implementation of Class-balanced Grouping and Sampling for Point Cloud 3D Object Detection.

Classes

CBGSDataset(dataset, class_map[, ignore])

Balance the number of scenes under different classes.

class CBGSDataset(dataset, class_map, ignore=-1)[source]

Balance the number of scenes under different classes.

__init__(dataset, class_map, ignore=-1)[source]: Creates an instance of the class.

__len__()[source]

Return the length of sample indices.

Returns:: Length of sample indices.
Return type:: int

__getitem__(idx)[source]

Get original dataset idx according to the given index.

Parameters:: idx (int) – The index of self.sample_indices.
Returns:: Data of the corresponding index.
Return type:: DictDataOrList