kaolin.datasets.scannet

class ScanNet(root_dir: str, scene_id: str, mode: Optional[str] = 'inference', transform: Optional[Callable] = None, label_transform: Optional[Callable] = None, loader: Optional[Callable] = None, color_mean: Optional[list] = [0.0, 0.0, 0.0], color_std: Optional[list] = [1.0, 1.0, 1.0], load_depth: Optional[bool] = False, seg_classes: Optional[str] = 'nyu40')[source]

ScanNet dataset http://www.scan-net.org/

Parameters

  • root_dir (str) – Path to the base directory of the dataset.

  • scene_file (str) – Path to file containing a list of scenes to be loaded.

  • transform (callable, optional) – A function/transform that takes in a PIL image and returns a transformed version of the image (default: None).

  • label_transform (callable, optional) – A function/transform that takes in the target and transforms it. (default: None).

  • loader (callable, optional) – A function to load an image given its path. By default, default_loader is used.

  • color_mean (list) – A list of length 3, containing the R, G, B channelwise mean.

  • color_std (list) – A list of length 3, containing the R, G, B channelwise standard deviation.

  • load_depth (bool) – Whether or not to load depth images (architectures that use depth information need depth to be loaded).

  • seg_classes (string) – The palette of classes that the network should learn.