org.datavec.image.loader

Class ImageLoader

java.lang.Object

org.datavec.image.loader.BaseImageLoader

org.datavec.image.loader.ImageLoader

All Implemented Interfaces:

Serializable

public class ImageLoader
extends BaseImageLoader

See Also:

Serialized Form

Nested Class Summary

Nested classes/interfaces inherited from class org.datavec.image.loader.BaseImageLoader

BaseImageLoader.MultiPageMode

Field Summary

Fields inherited from class org.datavec.image.loader.BaseImageLoader

ALLOWED_FORMATS, BASE_DIR, centerCropIfNeeded, channels, height, imageTransform, multiPageMode, rng, width

Constructor Summary

Constructors

Constructor and Description
ImageLoader()
ImageLoader(long height, long width) Instantiate an image with the given height and width
ImageLoader(long height, long width, long channels) Instantiate an image with the given height and width
ImageLoader(long height, long width, long channels, boolean centerCropIfNeeded) Instantiate an image with the given height and width

Method Summary

Modifier and Type	Method and Description
Image	asImageMatrix(File f) As per BaseImageLoader.asMatrix(File) but as an Image
Image	asImageMatrix(File f, boolean nchw) As per BaseImageLoader.asMatrix(File, boolean) but as an Image
Image	asImageMatrix(InputStream inputStream) As per BaseImageLoader.asMatrix(InputStream) but as an Image
Image	asImageMatrix(InputStream inputStream, boolean nchw) As per BaseImageLoader.asMatrix(InputStream, boolean) but as an Image
INDArray	asImageMiniBatches(File f, int numMiniBatches, int numRowsPerSlice) Slices up an image in to a mini batch.
INDArray	asMatrix(BufferedImage image) Convert an BufferedImage to a matrix
INDArray	asMatrix(File f) Convert an image file in to a matrix
INDArray	asMatrix(File f, boolean nchw) Load an image from a file to an INDArray
INDArray	asMatrix(InputStream inputStream) Convert an input stream to a matrix
INDArray	asMatrix(InputStream inputStream, boolean nchw) Load an image file from an input stream to an INDArray
INDArray	asRowVector(BufferedImage image) Convert an image in to a row vector
INDArray	asRowVector(File f) Convert a file to a row vector
INDArray	asRowVector(InputStream inputStream)
BufferedImage	centerCropIfNeeded(BufferedImage img)
int[]	flattenedImageFromFile(File f)
int[][]	fromFile(File file) Load a rastered image from file
int[][][]	fromFileMultipleChannels(File file) Load a rastered image from file
protected BufferedImage	scalingIfNeed(BufferedImage image, boolean needAlpha)
protected BufferedImage	scalingIfNeed(BufferedImage image, long dstHeight, long dstWidth, long dstImageType, boolean needAlpha)
INDArray	toBgr(BufferedImage image) Convert an BufferedImage to an bgr spectrum image
INDArray	toBgr(File file) Convert an input stream to an bgr spectrum image
INDArray	toBgr(InputStream inputStream) Convert an input stream to an bgr spectrum image
static BufferedImage	toBufferedImage(Image img, int type) Converts a given Image into a BufferedImage
void	toBufferedImageRGB(INDArray arr, BufferedImage image) Convert the given image to an rgb image
static BufferedImage	toImage(INDArray matrix) Convert a matrix in to a buffereed image
protected INDArray	toINDArrayBGR(BufferedImage image)
protected int[][]	toIntArrayArray(BufferedImage image)
INDArray	toRaveledTensor(BufferedImage image) Convert an image in to a raveled tensor of the bgr values of the image
INDArray	toRaveledTensor(File file) Changes the input stream in to an bgr based raveled(flattened) vector
INDArray	toRaveledTensor(InputStream is) Changes the input stream in to an bgr based raveled(flattened) vector

Methods inherited from class org.datavec.image.loader.BaseImageLoader

downloadAndUntar, getAllowedFormats

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

ImageLoader

public ImageLoader()

ImageLoader

public ImageLoader(long height,
                   long width)

Instantiate an image with the given height and width

Parameters:

height - the height to load*

width - the width to load

ImageLoader

public ImageLoader(long height,
                   long width,
                   long channels)

Instantiate an image with the given height and width

Parameters:

height - the height to load

width - the width to load

channels - the number of channels for the image*

ImageLoader

public ImageLoader(long height,
                   long width,
                   long channels,
                   boolean centerCropIfNeeded)

Instantiate an image with the given height and width

Parameters:

height - the height to load

width - the width to load

channels - the number of channels for the image*

centerCropIfNeeded - to crop before rescaling and converting

Method Detail

asRowVector

public INDArray asRowVector(File f)
                     throws IOException

Convert a file to a row vector

Specified by:

asRowVector in class BaseImageLoader

Parameters:

f - the image to convert

Returns:

the flattened image

Throws:

IOException

asRowVector

public INDArray asRowVector(InputStream inputStream)
                     throws IOException

Specified by:

asRowVector in class BaseImageLoader

Throws:

IOException

asRowVector

public INDArray asRowVector(BufferedImage image)

Convert an image in to a row vector

Parameters:

image - the image to convert

Returns:

the row vector based on a rastered representation of the image

toRaveledTensor

public INDArray toRaveledTensor(File file)

Changes the input stream in to an bgr based raveled(flattened) vector

Parameters:

file - the input stream to convert

Returns:

the raveled bgr values for this input stream

toRaveledTensor

public INDArray toRaveledTensor(InputStream is)

Changes the input stream in to an bgr based raveled(flattened) vector

Parameters:

is - the input stream to convert

Returns:

the raveled bgr values for this input stream

toRaveledTensor

public INDArray toRaveledTensor(BufferedImage image)

Convert an image in to a raveled tensor of the bgr values of the image

Parameters:

image - the image to parse

Returns:

the raveled tensor of bgr values

toBgr

public INDArray toBgr(File file)

Convert an input stream to an bgr spectrum image

Parameters:

file - the file to convert

Returns:

the input stream to convert

toBgr

public INDArray toBgr(InputStream inputStream)

Convert an input stream to an bgr spectrum image

Parameters:

inputStream - the input stream to convert

Returns:

the input stream to convert

toBgr

public INDArray toBgr(BufferedImage image)

Convert an BufferedImage to an bgr spectrum image

Parameters:

image - the BufferedImage to convert

Returns:

the input stream to convert

asMatrix

public INDArray asMatrix(File f)
                  throws IOException

Convert an image file in to a matrix

Specified by:

asMatrix in class BaseImageLoader

Parameters:

f - the file to convert

Returns:

a 2d matrix of a rastered version of the image

Throws:

IOException

asMatrix

public INDArray asMatrix(File f,
                         boolean nchw)
                  throws IOException

Description copied from class: BaseImageLoader

Load an image from a file to an INDArray

Specified by:

asMatrix in class BaseImageLoader

Parameters:

f - File to load the image from

nchw - If true: return image in NCHW/channels_first [1, channels, height width] format; if false, return in NHWC/channels_last [1, height, width, channels] format

Returns:

Image file as as INDArray

Throws:

IOException

asMatrix

public INDArray asMatrix(InputStream inputStream)
                  throws IOException

Convert an input stream to a matrix

Specified by:

asMatrix in class BaseImageLoader

Parameters:

inputStream - the input stream to convert

Returns:

the input stream to convert

Throws:

IOException

asMatrix

public INDArray asMatrix(InputStream inputStream,
                         boolean nchw)
                  throws IOException

Description copied from class: BaseImageLoader

Load an image file from an input stream to an INDArray

Specified by:

asMatrix in class BaseImageLoader

Parameters:

inputStream - Input stream to load the image from

nchw - If true: return image in NCHW/channels_first [1, channels, height width] format; if false, return in NHWC/channels_last [1, height, width, channels] format

Returns:

Image file stream as as INDArray

Throws:

IOException

asImageMatrix

public Image asImageMatrix(File f)
                    throws IOException

Description copied from class: BaseImageLoader

As per BaseImageLoader.asMatrix(File) but as an Image

Specified by:

asImageMatrix in class BaseImageLoader

Throws:

IOException

asImageMatrix

public Image asImageMatrix(File f,
                           boolean nchw)
                    throws IOException

Description copied from class: BaseImageLoader

As per BaseImageLoader.asMatrix(File, boolean) but as an Image

Specified by:

asImageMatrix in class BaseImageLoader

Throws:

IOException

asImageMatrix

public Image asImageMatrix(InputStream inputStream)
                    throws IOException

Description copied from class: BaseImageLoader

As per BaseImageLoader.asMatrix(InputStream) but as an Image

Specified by:

asImageMatrix in class BaseImageLoader

Throws:

IOException

asImageMatrix

public Image asImageMatrix(InputStream inputStream,
                           boolean nchw)
                    throws IOException

Description copied from class: BaseImageLoader

As per BaseImageLoader.asMatrix(InputStream, boolean) but as an Image

Specified by:

asImageMatrix in class BaseImageLoader

Throws:

IOException

asMatrix

public INDArray asMatrix(BufferedImage image)

Convert an BufferedImage to a matrix

Parameters:

image - the BufferedImage to convert

Returns:

the input stream to convert

asImageMiniBatches

public INDArray asImageMiniBatches(File f,
                                   int numMiniBatches,
                                   int numRowsPerSlice)

Slices up an image in to a mini batch.

Parameters:

f - the file to load from

numMiniBatches - the number of images in a mini batch

numRowsPerSlice - the number of rows for each image

Returns:

a tensor representing one image as a mini batch

flattenedImageFromFile

public int[] flattenedImageFromFile(File f)
                             throws IOException

Throws:

IOException

fromFile

public int[][] fromFile(File file)
                 throws IOException

Load a rastered image from file

Parameters:

file - the file to load

Returns:

the rastered image

Throws:

IOException

fromFileMultipleChannels

public int[][][] fromFileMultipleChannels(File file)
                                   throws IOException

Load a rastered image from file

Parameters:

file - the file to load

Returns:

the rastered image

Throws:

IOException

toImage

public static BufferedImage toImage(INDArray matrix)

Convert a matrix in to a buffereed image

Parameters:

matrix - the

Returns:

BufferedImage

toBufferedImageRGB

public void toBufferedImageRGB(INDArray arr,
                               BufferedImage image)

Convert the given image to an rgb image

Parameters:

arr - the array to use

image - the image to set

toBufferedImage

public static BufferedImage toBufferedImage(Image img,
                                            int type)

Converts a given Image into a BufferedImage

Parameters:

img - The Image to be converted

type - The color model of BufferedImage

Returns:

The converted BufferedImage

toIntArrayArray

protected int[][] toIntArrayArray(BufferedImage image)

toINDArrayBGR

protected INDArray toINDArrayBGR(BufferedImage image)

centerCropIfNeeded

public BufferedImage centerCropIfNeeded(BufferedImage img)

scalingIfNeed

protected BufferedImage scalingIfNeed(BufferedImage image,
                                      boolean needAlpha)

scalingIfNeed

protected BufferedImage scalingIfNeed(BufferedImage image,
                                      long dstHeight,
                                      long dstWidth,
                                      long dstImageType,
                                      boolean needAlpha)