DocumentPreprocessor<T>

Models & Types · AiDotNet.Preprocessing.Document

DocumentPreprocessor - Comprehensive document image preprocessing pipeline.

Implements: IDisposable

For Beginners

Document preprocessing improves model accuracy by:

  • Normalizing image characteristics
  • Removing noise and artifacts
  • Correcting geometric distortions
  • Enhancing text contrast

Key features:

  • Chained preprocessing pipeline
  • Configurable operation order
  • Quality-aware preprocessing
  • Batch processing support

Example usage:

How It Works

DocumentPreprocessor provides a unified interface for applying multiple preprocessing operations to document images before feeding them to AI models.

Constructors

Constructor Summary
DocumentPreprocessor() Creates a DocumentPreprocessor with default settings.

Methods

Method Summary
IList<Tensor<T>> PreprocessBatch(IEnumerable<Tensor<T>> images, DocumentPreprocessingOptions options) Applies preprocessing to multiple document images.
Tensor<T> CenterCrop(Tensor<T> image, int targetWidth, int targetHeight) Center crops an image to the specified dimensions.
Tensor<T> Pad(Tensor<T> image, int targetWidth, int targetHeight, T paddingValue) Pads an image to the specified dimensions.
Tensor<T> Preprocess(Tensor<T> image, DocumentPreprocessingOptions options) Applies the full preprocessing pipeline to a document image.
Tensor<T> Resize(Tensor<T> image, int targetWidth, int targetHeight, InterpolationMethod method) Resizes an image to the specified dimensions.
Tensor<T> ToGrayscale(Tensor<T> image) Converts an RGB image to grayscale.
void Dispose()