Module NewOCR

Package com.uddernetworks.newocr.utils

Class OCRUtils

java.lang.Object

com.uddernetworks.newocr.utils.OCRUtils

public class OCRUtils
extends java.lang.Object

Some various utility methods used by the OCR that may assist others using the library.

Since:

April 25, 2019

Constructor Summary

Constructors

Constructor	Description
OCRUtils()

Method Summary

Modifier and Type	Method	Description
static void	colorColumn(java.awt.image.BufferedImage image, java.awt.Color color, int x, int y, int height)	Colors in a vertical line of an image.
static void	colorRow(java.awt.image.BufferedImage image, java.awt.Color color, int y, int x, int width)	Colors in a horizontal line of an image.
static boolean[][]	createGrid(java.awt.image.BufferedImage bufferedImage)	Creates a grid of booleans from a BufferedImage with the same dimensions as the image.
static double	diff(double one, double two)	Gets the difference between two doubles.
static int	diff(int one, int two)	Gets the difference between two ints
static java.util.Optional<java.awt.image.BufferedImage>	filter(java.awt.image.BufferedImage input)	Binarizes the input image, making all pixels wither black or white with an alpha of 255
static java.util.OptionalDouble	getDifferencesFrom(double[] input1, double[] input2)	Gets the difference of two arrays' values.
static boolean	isBlack(java.awt.image.BufferedImage image, int x, int y)	Gets if a pixel should be considered black.
static boolean	isRowPopulated(boolean[][] values, int y)	Gets if the row has any `true` (Black) values in it
static boolean	isWithin(double lowerBound, double upperBound, double value)	Gets if a given number is within two bounds.
static void	makeImage(boolean[][] values, java.lang.String path)	Creates an image from a grid of booleans, `true` being black and `false` being white.
static void	printOut(boolean[][] values)	Prints a grid of booleans to console using full width characters so it will appear proportional and not skewed with spaces and the filling character.
static java.awt.image.BufferedImage	readImage(java.io.File input)	An ImageIO.read() replacement, which in tests can be up to 15x faster.
static java.lang.String	removeLeadingSpaces(java.lang.String string)	Removes all common spaces between all newlines, useful if the OCR say adds an extra 2 spaces before all lines of text, this will remove the 2 spaces.
static void	toGrid(java.awt.image.BufferedImage input, boolean[][] values)	Populates a boolean 2D array with the same dimensions as the input image where each pixel is represented by a boolean value, black being `true`, white being `false`.

Methods inherited from class java.lang.Object

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

Constructor Detail

OCRUtils

public OCRUtils()

Method Detail

readImage

public static java.awt.image.BufferedImage readImage(java.io.File input)
                                              throws java.io.IOException

An ImageIO.read() replacement, which in tests can be up to 15x faster. This has shown to significantly improve the OCR's performance both in training and actual usage.

Parameters:

input - The file to read

Returns:

The BufferedImage of the file

Throws:

java.io.IOException - If there are issues reading the file

removeLeadingSpaces

public static java.lang.String removeLeadingSpaces(java.lang.String string)

Removes all common spaces between all newlines, useful if the OCR say adds an extra 2 spaces before all lines of text, this will remove the 2 spaces.

Parameters:

string - The input string

Returns:

The input string trimmed properly

diff

public static double diff(double one,
                          double two)

Gets the difference between two doubles.

Parameters:

one - The first number

two - The second number

Returns:

The difference

diff

public static int diff(int one,
                       int two)

Gets the difference between two ints

Parameters:

one - The first number

two - The second number

Returns:

The difference

getDifferencesFrom

public static java.util.OptionalDouble getDifferencesFrom(double[] input1,
                                                          double[] input2)

Gets the difference of two arrays' values.

Parameters:

input1 - The first array

input2 - The second array

Returns:

An array with the same length as the inputs containing the difference of both arrays' respective values

isWithin

public static boolean isWithin(double lowerBound,
                               double upperBound,
                               double value)

Gets if a given number is within two bounds.

Parameters:

lowerBound - The lower bound to check

upperBound - The upper bound to check

value - The value to check

Returns:

If the two values are within the given bounds

createGrid

public static boolean[][] createGrid(java.awt.image.BufferedImage bufferedImage)

Creates a grid of booleans from a BufferedImage with the same dimensions as the image.

Parameters:

bufferedImage - The input BufferedImage

Returns:

The created grid

toGrid

public static void toGrid(java.awt.image.BufferedImage input,
                          boolean[][] values)

Populates a boolean 2D array with the same dimensions as the input image where each pixel is represented by a boolean value, black being `true`, white being `false`.

Parameters:

input - The input image

values - The mutable empty grid

isRowPopulated

public static boolean isRowPopulated(boolean[][] values,
                                     int y)

Gets if the row has any `true` (Black) values in it

Parameters:

values - The grid of image values

y - The Y coordinate of the row to check

Returns:

If the row has anything in it

filter

public static java.util.Optional<java.awt.image.BufferedImage> filter(java.awt.image.BufferedImage input)

Binarizes the input image, making all pixels wither black or white with an alpha of 255

Parameters:

input - The input image to be filtered

Returns:

The filtered image

isBlack

public static boolean isBlack(java.awt.image.BufferedImage image,
                              int x,
                              int y)

Gets if a pixel should be considered black.

Parameters:

image - The input image

x - The X coordinate to check

y - The Y coordinate to check

Returns:

If the pixel should be considered black

makeImage

public static void makeImage(boolean[][] values,
                             java.lang.String path)

Creates an image from a grid of booleans, `true` being black and `false` being white.

Parameters:

values - The values to convert into an image

path - The path of the file

colorRow

public static void colorRow(java.awt.image.BufferedImage image,
                            java.awt.Color color,
                            int y,
                            int x,
                            int width)

Colors in a horizontal line of an image.

Parameters:

image - The image to color on

color - The color to use

y - The Y value of the horizontal line

x - The X start position of the line

width - The width of the line to draw

colorColumn

public static void colorColumn(java.awt.image.BufferedImage image,
                               java.awt.Color color,
                               int x,
                               int y,
                               int height)

Colors in a vertical line of an image.

Parameters:

image - The image to color on

color - The color to use

y - The Y start position of the line

x - The X value of the vertical line

height - The height of the line to draw

printOut

public static void printOut(boolean[][] values)

Prints a grid of booleans to console using full width characters so it will appear proportional and not skewed with spaces and the filling character.

Parameters:

values - The values to print out