ScanNetSemanticSegmentationDataLoaderOptions

Options & Configuration · AiDotNet.Data.Geometry

Configuration options for the ScanNet semantic segmentation data loader.

Constructors

Constructor Summary
ScanNetSemanticSegmentationDataLoaderOptions() Initializes a new instance with default values.

Properties

Property Summary
DatasetSplit Split { get; set; } Dataset split to load.
PointPaddingStrategy PaddingStrategy { get; set; } Strategy for padding when fewer points exist than requested.
PointSamplingStrategy SamplingStrategy { get; set; } Strategy for sampling points from each scene.
ScanNetInputFormat InputFormat { get; set; } Input data format selection.
ScanNetLabelMode LabelMode { get; set; } Label mapping mode.
bool AutoDetectLabelColumn { get; set; } Whether to auto-detect a label column in preprocessed text files.
bool AutoDownload { get; set; } Automatically download the dataset if not present.
bool IncludeColors { get; set; } Whether to include RGB colors when available.
bool IncludeNormals { get; set; } Whether to include normals when available.
bool IncludeUnknownClass { get; set; } Whether to reserve an explicit unknown class at index 0.
bool NormalizeColors { get; set; } Whether to normalize colors from 0-255 to 0-1.
int PointsPerSample { get; set; } Number of points per sample.
int? MaxSamples { get; set; } Optional cap on the number of samples to load.
int? RandomSeed { get; set; } Optional random seed for reproducible sampling.
string DataPath { get; set; } Root data path.