ImagePatchSplitter<T>
Models & Types · AiDotNet.Preprocessing.DataPreparation.Splitting.DomainSpecific
Image patch splitter for image segmentation and patch-based learning tasks.
Inheritance: DataSplitterBase<T> → ImagePatchSplitter<T>
For Beginners
In image analysis, we often work with patches (small regions) extracted from larger images. Adjacent patches can be highly correlated, so we need to ensure patches from the same region don’t appear in both train and test.
How It Works
How It Works:
- Assume each sample is a patch with (imageId, x, y) metadata
- Group patches by source image
- Either split entire images, or split with spatial buffer
When to Use:
- Medical image segmentation
- Satellite/aerial image analysis
- Microscopy image analysis
- Any patch-based computer vision task
Constructors
| Constructor | Summary |
|---|---|
ImagePatchSplitter(double testSize, int patchSize, int overlap, int imageIdColumn, bool shuffle, int randomSeed) |
Creates a new image patch splitter. |
Properties
| Property | Summary |
|---|---|
string Description { get; } |
Methods
| Method | Summary |
|---|---|
DataSplitResult<T> Split(Matrix<T> X, Vector<T> y) |