Add File
This commit is contained in:
302
src/main/java/org/dromara/easyai/entity/RGBNorm.java
Normal file
302
src/main/java/org/dromara/easyai/entity/RGBNorm.java
Normal file
@@ -0,0 +1,302 @@
|
||||
package org.dromara.easyai.entity;
|
||||
|
||||
import org.dromara.easyai.matrixTools.Matrix;
|
||||
import org.dromara.easyai.tools.RgbRegression;
|
||||
|
||||
import java.util.*;
|
||||
|
||||
public class RGBNorm {
|
||||
private float[] rgbAll;
|
||||
private float norm;
|
||||
private int nub;
|
||||
private float[] rgb;//均值
|
||||
private float[] rgbUp;
|
||||
private List<float[]> rgbs = new ArrayList<>();//需要对它进行排序
|
||||
private List<Float> powers = new ArrayList<>();//需要对它进行排序
|
||||
private RgbRegression rgbRegression;
|
||||
private int len;
|
||||
private float varAll = 1;
|
||||
private Matrix avgMatrix;//均值矩阵
|
||||
private Matrix varMatrix;//方差矩阵
|
||||
private float gmParameter;//gm混合系数
|
||||
private float probabilitySigma = 0;//后验概率求和
|
||||
|
||||
public float getVarAll() {
|
||||
return varAll;
|
||||
}
|
||||
|
||||
public Matrix getAvgMatrix() {
|
||||
return avgMatrix;
|
||||
}
|
||||
|
||||
public Matrix getVarMatrix() {
|
||||
return varMatrix;
|
||||
}
|
||||
|
||||
public List<float[]> getRgbs() {
|
||||
return rgbs;
|
||||
}
|
||||
|
||||
public RgbRegression getRgbRegression() {
|
||||
return rgbRegression;
|
||||
}
|
||||
|
||||
public void setRgbRegression(RgbRegression rgbRegression) {
|
||||
this.rgbRegression = rgbRegression;
|
||||
}
|
||||
|
||||
public RGBNorm(float[] rgb, int len) {
|
||||
this.len = len;
|
||||
rgbAll = new float[len];
|
||||
this.rgb = new float[len];
|
||||
this.rgbUp = rgb;
|
||||
gmParameter = new Random().nextFloat();
|
||||
}
|
||||
|
||||
public RGBNorm() {
|
||||
}
|
||||
|
||||
public void syn() {
|
||||
rgbUp = rgb;
|
||||
}
|
||||
|
||||
public void clearRGB() {
|
||||
probabilitySigma = 0;
|
||||
rgbs.clear();
|
||||
powers.clear();
|
||||
}
|
||||
|
||||
//特征及权重
|
||||
public void setGmFeature(float[] feature, float probability) {
|
||||
rgbs.add(feature);
|
||||
powers.add(probability);
|
||||
probabilitySigma = probability + probabilitySigma;
|
||||
}
|
||||
|
||||
public void clear() {
|
||||
rgbAll = new float[len];
|
||||
nub = 0;
|
||||
for (int i = 0; i < rgb.length; i++) {
|
||||
rgbUp[i] = rgb[i];
|
||||
}
|
||||
rgbs.clear();
|
||||
//System.out.println("clear==" + Arrays.toString(rgbUp));
|
||||
}
|
||||
|
||||
public int getNub() {
|
||||
return nub;
|
||||
}
|
||||
|
||||
public boolean compare() {
|
||||
boolean isNext = false;
|
||||
for (int i = 0; i < rgb.length; i++) {
|
||||
if (rgb[i] != rgbUp[i]) {
|
||||
isNext = true;
|
||||
break;
|
||||
}
|
||||
}
|
||||
return isNext;
|
||||
}
|
||||
|
||||
public float getEDist(float[] x) {
|
||||
float[] y = new float[x.length];
|
||||
for (int i = 0; i < y.length; i++) {
|
||||
y[i] = x[i] - rgbUp[i];
|
||||
}
|
||||
float sigma = 0;
|
||||
for (int i = 0; i < y.length; i++) {
|
||||
sigma = sigma + (float)Math.pow(y[i], 2);
|
||||
}
|
||||
return (float)Math.sqrt(sigma);
|
||||
}
|
||||
|
||||
public void setColor(float[] rgb) {
|
||||
for (int i = 0; i < rgb.length; i++) {
|
||||
rgbAll[i] = rgbAll[i] + rgb[i];
|
||||
}
|
||||
rgbs.add(rgb);
|
||||
nub++;
|
||||
}
|
||||
|
||||
private Matrix average(Matrix matrix) throws Exception {//计算平均值
|
||||
int x = matrix.getX();
|
||||
int y = matrix.getY();
|
||||
Matrix avgMatrix = new Matrix(x, 1);
|
||||
boolean isPower = powers.size() > 1;
|
||||
for (int i = 0; i < x; i++) {
|
||||
for (int j = 0; j < y; j++) {
|
||||
float power;
|
||||
if (!isPower) {
|
||||
power = 1;
|
||||
} else {
|
||||
power = powers.get(j);
|
||||
}
|
||||
float nub = avgMatrix.getNumber(i, 0) + matrix.getNumber(i, j) * power;
|
||||
avgMatrix.setNub(i, 0, nub);
|
||||
}
|
||||
}
|
||||
for (int i = 0; i < x; i++) {
|
||||
float nub;
|
||||
if (probabilitySigma > 0) {
|
||||
nub = avgMatrix.getNumber(i, 0) / probabilitySigma;
|
||||
} else {
|
||||
nub = avgMatrix.getNumber(i, 0) / y;
|
||||
}
|
||||
avgMatrix.setNub(i, 0, nub);
|
||||
}
|
||||
return avgMatrix;
|
||||
}
|
||||
|
||||
private Matrix getVariance(Matrix matrix, Matrix avgMatrix) throws Exception {
|
||||
int x = matrix.getX();
|
||||
int y = matrix.getY();
|
||||
Matrix varMatrix = new Matrix(x, 1);
|
||||
boolean isPower = powers.size() > 1;
|
||||
for (int i = 0; i < x; i++) {
|
||||
float avg = avgMatrix.getNumber(i, 0);
|
||||
for (int j = 0; j < y; j++) {
|
||||
float power;
|
||||
if (!isPower) {
|
||||
power = 1;
|
||||
} else {
|
||||
power = powers.get(j);
|
||||
}
|
||||
float sub = (float) ((float)Math.pow(matrix.getNumber(i, j) - avg, 2) * power);
|
||||
varMatrix.setNub(i, 0, sub + varMatrix.getNumber(i, 0));
|
||||
}
|
||||
}
|
||||
for (int i = 0; i < x; i++) {
|
||||
float nub;
|
||||
if (probabilitySigma > 0) {
|
||||
nub = varMatrix.getNumber(i, 0) / probabilitySigma;
|
||||
} else {
|
||||
nub = varMatrix.getNumber(i, 0) / y;
|
||||
}
|
||||
varMatrix.setNub(i, 0, nub);
|
||||
}
|
||||
return varMatrix;
|
||||
}
|
||||
|
||||
public float[] getFeature() throws Exception {
|
||||
int avgLen = avgMatrix.getX();
|
||||
int length = avgLen * 2 + 1;
|
||||
float[] feature = new float[length];
|
||||
for (int i = 0; i < length - 1; i++) {
|
||||
if (i < avgLen) {
|
||||
feature[i] = avgMatrix.getNumber(i, 0);
|
||||
} else {
|
||||
int t = i - avgLen;
|
||||
feature[i] = varMatrix.getNumber(t, 0);
|
||||
}
|
||||
}
|
||||
feature[length - 1] = gmParameter;
|
||||
return feature;
|
||||
}
|
||||
|
||||
public void insertFeature(float[] feature) throws Exception {
|
||||
int length = feature.length - 1;
|
||||
int size = length / 2;
|
||||
avgMatrix = new Matrix(size, 1);
|
||||
varMatrix = new Matrix(size, 1);
|
||||
gmParameter = feature[length];
|
||||
varAll = 1;
|
||||
for (int i = 0; i < length; i++) {
|
||||
if (i < size) {
|
||||
avgMatrix.setNub(i, 0, feature[i]);
|
||||
} else {
|
||||
int t = i - size;
|
||||
float var = feature[i];
|
||||
varMatrix.setNub(t, 0, var);
|
||||
varAll = varAll * (float)Math.sqrt(var);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
public void gm() throws Exception {//混合高斯聚类模型参数计算
|
||||
int size = rgbs.size();
|
||||
if (size > 0) {
|
||||
Matrix matrix = new Matrix(len, size);
|
||||
for (int i = 0; i < size; i++) {
|
||||
float[] rgb = rgbs.get(i);
|
||||
for (int j = 0; j < len; j++) {
|
||||
matrix.setNub(j, i, rgb[j]);
|
||||
}
|
||||
}
|
||||
avgMatrix = average(matrix);//均值矩阵
|
||||
varMatrix = getVariance(matrix, avgMatrix);//方差矩阵
|
||||
varAll = 1;
|
||||
for (int i = 0; i < len; i++) {
|
||||
float var = (float)Math.sqrt(varMatrix.getNumber(i, 0));
|
||||
varAll = varAll * var;
|
||||
}
|
||||
//确认新的混合系数
|
||||
if (probabilitySigma > 0) {
|
||||
gmParameter = probabilitySigma / rgbs.size();
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
public float getGMProbability(float[] feature) throws Exception {//计算正态分布概率密度
|
||||
float zSigma = 0;
|
||||
int size = feature.length;
|
||||
for (int i = 0; i < size; i++) {
|
||||
float sub = (float)Math.pow(feature[i] - avgMatrix.getNumber(i, 0), 2) / varMatrix.getNumber(i, 0);
|
||||
zSigma = sub + zSigma;
|
||||
}
|
||||
float one = 1 / ((float)Math.pow((float)Math.sqrt((float)Math.PI * 2), size) * varAll);
|
||||
float two = (float)Math.exp(zSigma * -0.5);
|
||||
float n = one * two * gmParameter;
|
||||
return n;
|
||||
}
|
||||
|
||||
public float getGmParameter() {
|
||||
return gmParameter;
|
||||
}
|
||||
|
||||
public void setGmParameter(float gmParameter) {
|
||||
this.gmParameter = gmParameter;
|
||||
}
|
||||
|
||||
public void norm() {//范长计算
|
||||
float sigma = 0;
|
||||
if (nub > 0) {
|
||||
for (int i = 0; i < rgb.length; i++) {
|
||||
float rgbc = rgbAll[i] / nub;
|
||||
rgb[i] = rgbc;
|
||||
sigma = sigma + (float)Math.pow(rgbc, 2);
|
||||
}
|
||||
norm = (float)Math.sqrt(sigma);
|
||||
}
|
||||
}
|
||||
|
||||
public void finish() {//进行排序
|
||||
RGBListSort rgbListSort = new RGBListSort();
|
||||
Collections.sort(rgbs, rgbListSort);
|
||||
}
|
||||
|
||||
public float getNorm() {
|
||||
return norm;
|
||||
}
|
||||
|
||||
public float[] getRgb() {
|
||||
return rgb;
|
||||
}
|
||||
|
||||
class RGBListSort implements Comparator<float[]> {
|
||||
@Override
|
||||
public int compare(float[] o1, float[] o2) {
|
||||
float o1Norm = 0;
|
||||
float o2Norm = 0;
|
||||
for (int i = 0; i < o1.length; i++) {
|
||||
o1Norm = o1Norm + (float)Math.pow(o1[i], 2);
|
||||
o2Norm = o2Norm + (float)Math.pow(o2[i], 2);
|
||||
}
|
||||
if (o1Norm > o2Norm) {
|
||||
return 1;
|
||||
} else if (o1Norm < o2Norm) {
|
||||
return -1;
|
||||
}
|
||||
return 0;
|
||||
}
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user