tool/io.github.kmplitert.tool.image/LiteRtImage

LiteRtImage

expect class LiteRtImage

Represents an image that can be preprocessed before being fed into a LiteRT model.

This class provides common image operations such as resizing and converting pixel data into a normalized FloatArray.

Platform-specific implementations are provided for each supported target.

Types

Companion

expect object Companion

Properties

channels

expect val channels: Int

The number of color channels in the image (e.g., 3 for RGB, 4 for RGBA, 1 for Grayscale).

height

expect val height: Int

The height of the image in pixels.

width

expect val width: Int

The width of the image in pixels.

Functions

centerCrop

expect fun centerCrop(width: Int, height: Int): LiteRtImage

Returns a center-cropped copy of this image.

crop

expect fun crop(x: Int, y: Int, width: Int, height: Int): LiteRtImage

Returns a cropped copy of this image.

flip

expect fun flip(horizontal: Boolean, vertical: Boolean): LiteRtImage

Returns a flipped copy of this image.

resize

expect fun resize(width: Int, height: Int): LiteRtImage

Returns a resized copy of this image.

rotate

expect fun rotate(degrees: Float): LiteRtImage

Returns a rotated copy of this image.

toBooleanArray

expect fun toBooleanArray(): BooleanArray

Converts the image into a boolean array.

toFloatArray

expect fun toFloatArray(mean: Float = 0.0f, std: Float = 1.0f): FloatArray

Converts the image into a normalized float array.

toGrayscale

expect fun toGrayscale(): LiteRtImage

Converts the image to grayscale.

toInt8Array

expect fun toInt8Array(): ByteArray

Converts the image into a signed 8-bit integer array (ByteArray).

toIntArray

expect fun toIntArray(): IntArray

Converts the image into an integer array.

toLongArray

expect fun toLongArray(): LongArray

Converts the image into a long array.

toRgb

expect fun toRgb(): LiteRtImage

Converts the image to RGB format (discarding alpha if present).

writeFloatBuffer

expect fun writeFloatBuffer(buffer: TFBuffer, mean: Float = 0.0f, std: Float = 1.0f)

Directly writes the image pixel data to a float TFBuffer with normalization.

writeInt8Buffer

expect fun writeInt8Buffer(buffer: TFBuffer)

Directly writes the image pixel data to a signed 8-bit integer io.github.kmplitert.core.TFBuffer.