com.htmlhifive.pitalium.image.util

クラス SimilarityUtils

java.lang.Object

com.htmlhifive.pitalium.image.util.SimilarityUtils

public class SimilarityUtils
extends java.lang.Object

Utility class to calculate similarity.

メソッドのサマリー

修飾子とタイプ	メソッドと説明
static double	calcFeatureDistance(java.awt.Color[][] expectedFeature, java.awt.Color[][] actualFeature) Calculate the distance of two feature matrices.
static SimilarityUnit	calcSimilarity(java.awt.image.BufferedImage expectedImage, java.awt.image.BufferedImage actualImage, java.awt.Rectangle rectangle, ComparedRectangleArea similarRectangle, Offset offset) calculate similarity of given rectangle area and offset And then, build similarRectangle using similarity values.
static SimilarityUnit	calcSimilarity(java.awt.image.BufferedImage expectedImage, java.awt.image.BufferedImage actualImage, java.awt.Rectangle rectangle, ComparedRectangleArea similarRectangle, Offset offset, double similarityFeatureMatrix) calculate similarity of given rectangle area and offset, but similarityFeatureMatrix is given.
static double	calcSimilarityByFeatureMatrix(java.awt.image.BufferedImage expectedImage, java.awt.image.BufferedImage actualImage, java.awt.Rectangle rectangle, Offset offset) Calculate the similarity using feature matrix and find the best match where it has the highest similarity This method should be implemented only when the size of actualSubImage is greater than or equal to FeatureCol by FeatureRow.
static double	calcSimilarityByFeatureMatrix(java.awt.image.BufferedImage expectedImage, java.awt.image.BufferedImage actualImage, java.awt.Rectangle expectedFrame, java.awt.Rectangle actualFrame) Calculate the similarity using feature matrix using fixed frames of expected and actual images.
static SimilarityUnit	calcSimilarityPixelByPixel(java.awt.image.BufferedImage expectedImage, java.awt.image.BufferedImage actualImage, java.awt.Rectangle rectangle, Offset offset) Calculate the similarity by comparing two images pixel by pixel, and find the best match where it has the highest similarity (when given offset is null).
static boolean	checkFeatureSize(int width, int height) Check if the size of rectangle is large enough to use feature method.

クラスから継承されたメソッド java.lang.Object

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

メソッドの詳細

calcSimilarity

public static SimilarityUnit calcSimilarity(java.awt.image.BufferedImage expectedImage,
                                            java.awt.image.BufferedImage actualImage,
                                            java.awt.Rectangle rectangle,
                                            ComparedRectangleArea similarRectangle,
                                            Offset offset)

calculate similarity of given rectangle area and offset And then, build similarRectangle using similarity values.

パラメータ:

expectedImage -

actualImage -

rectangle -

similarRectangle -

offset -

戻り値:

TODO

calcSimilarity

public static SimilarityUnit calcSimilarity(java.awt.image.BufferedImage expectedImage,
                                            java.awt.image.BufferedImage actualImage,
                                            java.awt.Rectangle rectangle,
                                            ComparedRectangleArea similarRectangle,
                                            Offset offset,
                                            double similarityFeatureMatrix)

calculate similarity of given rectangle area and offset, but similarityFeatureMatrix is given. And then, build similarRectangle using similarity values.

パラメータ:

expectedImage -

actualImage -

rectangle -

similarRectangle -

offset -

similarityFeatureMatrix - in the case of "SCALING", we use similarity already calculated. in the other cases, it has -1.

calcFeatureDistance

public static double calcFeatureDistance(java.awt.Color[][] expectedFeature,
                                         java.awt.Color[][] actualFeature)

Calculate the distance of two feature matrices.

パラメータ:

expectedFeature - The feature matrix of expected sub image

actualFeature - The feature matrix of actual sub image

戻り値:

the norm of distance between two matrices

checkFeatureSize

public static boolean checkFeatureSize(int width,
                                       int height)

Check if the size of rectangle is large enough to use feature method.

パラメータ:

width - the width of compared rectangle

heigth - the height of compared rectangle

戻り値:

if the rectangle is large enough, return true.

calcSimilarityByFeatureMatrix

public static double calcSimilarityByFeatureMatrix(java.awt.image.BufferedImage expectedImage,
                                                   java.awt.image.BufferedImage actualImage,
                                                   java.awt.Rectangle rectangle,
                                                   Offset offset)

Calculate the similarity using feature matrix and find the best match where it has the highest similarity This method should be implemented only when the size of actualSubImage is greater than or equal to FeatureCol by FeatureRow.

パラメータ:

expectedImage -

actualImage -

rectangle - The rectangle area to be compared of actual image.

戻り値:

the 'feature' similarity of given area between two images.

calcSimilarityByFeatureMatrix

public static double calcSimilarityByFeatureMatrix(java.awt.image.BufferedImage expectedImage,
                                                   java.awt.image.BufferedImage actualImage,
                                                   java.awt.Rectangle expectedFrame,
                                                   java.awt.Rectangle actualFrame)

Calculate the similarity using feature matrix using fixed frames of expected and actual images. This method should be implemented only when the size of actualSubImage is greater than or equal to FeatureCol by FeatureRow.

パラメータ:

expectedImage -

actualImage -

expectedFrame - the rectangle area to be extracted feature vector in expectedImage

actualFrame - the rectangle area to be extracted feature vector in actualImage

戻り値:

the 'feature' similarity of given area between two images.

calcSimilarityPixelByPixel

public static SimilarityUnit calcSimilarityPixelByPixel(java.awt.image.BufferedImage expectedImage,
                                                        java.awt.image.BufferedImage actualImage,
                                                        java.awt.Rectangle rectangle,
                                                        Offset offset)

Calculate the similarity by comparing two images pixel by pixel, and find the best match where it has the highest similarity (when given offset is null). In this method, we count the number of different pixels as well.

パラメータ:

expectedSubImage - the sub-image of given rectangle area of expected image

actualSubImage - the sub-image of given 'template' rectangle area of actual image. it is smaller than expectedSubImage.

rectangle - The rectangle area where to compare.

similarityUnit -

offset - best match offset. If default offset is given, don't find the best match.

戻り値: