Pad<T>

Models & Types · AiDotNet.Augmentation.Image

Pads an image with configurable padding amounts and fill modes.

Inheritance: AugmentationBase<T, ImageTensor<T>> → SpatialAugmentationBase<T, ImageTensor<T>> → SpatialImageAugmenterBase<T> → Pad<T>

For Beginners

Padding makes an image larger by adding extra pixels around the edges. This is useful when you need a specific size but don’t want to crop or stretch the image. Black padding (constant=0) is most common.

How It Works

Padding adds pixels around the border of an image. Multiple fill modes control how the new pixels are filled:

  • Constant: Fills with a fixed value (e.g., 0 for black, 255 for white)
  • Edge: Extends the edge pixel values outward
  • Reflect: Mirrors the image at the boundary
  • Symmetric: Like reflect but includes the boundary pixel

When to use:

  • Making images a specific size without distortion
  • Preparing for convolutions that reduce spatial size
  • Adding borders for visual presentation

Constructors

Constructor Summary
Pad(int padHorizontal, int padVertical, PaddingMode mode, double fillValue, double probability) Creates a new pad augmentation with separate horizontal and vertical padding.
Pad(int padLeft, int padRight, int padTop, int padBottom, PaddingMode mode, double fillValue, double probability) Creates a new pad augmentation with individual padding per side.
Pad(int padding, PaddingMode mode, double fillValue, double probability) Creates a new pad augmentation with uniform padding.

Properties

Property Summary
PaddingMode Mode { get; } Gets the padding mode.
double FillValue { get; } Gets the constant fill value (used when Mode is Constant).
int PadBottom { get; } Gets the padding on the bottom side.
int PadLeft { get; } Gets the padding on the left side.
int PadRight { get; } Gets the padding on the right side.
int PadTop { get; } Gets the padding on the top side.

Methods

Method Summary
IDictionary<string, object> GetParameters()