m53297601
/
easyAi
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

末尾加了一层池化，减轻矩阵求逆的压力

Browse Source
pull/1/head
lidapeng 5 years ago
parent b5ee5c9338
commit d8d6a3197c
5 changed files with 95 additions and 37 deletions
Show Stats Download Patch File Download Diff File

46
src/main/java/org/wlld/MatrixTools/MatrixOperation.java
Unescape View File

@ -107,6 +107,52 @@ public class MatrixOperation {
}
}
public static Matrix getPoolVector(Matrix matrix) throws Exception {
if (matrix.getX() == 1 || matrix.getY() == 1) {
Matrix vector;
int nub;
boolean isRow = false;
if (matrix.getX() == 1) {//行向量
isRow = true;
nub = matrix.getY() / 4;
vector = new Matrix(1, nub);
} else {//列向量
nub = matrix.getX() / 4;
vector = new Matrix(nub, 1);
}
int k = 0;
for (int i = 0; i < nub * 4 - 3; i += 4) {
double max = 0;
if (isRow) {
max = matrix.getNumber(0, i);
max = getMax(max, matrix.getNumber(0, i + 1));
max = getMax(max, matrix.getNumber(0, i + 2));
max = getMax(max, matrix.getNumber(0, i + 3));
vector.setNub(0, k, max);
} else {
max = matrix.getNumber(i, 0);
max = getMax(max, matrix.getNumber(i + 1, 0));
max = getMax(max, matrix.getNumber(i + 2, 0));
max = getMax(max, matrix.getNumber(i + 3, 0));
vector.setNub(k, 0, max);
}
k++;
}
return vector;
} else {
throw new Exception("this matrix is not a vector");
}
}
private static double getMax(double o1, double o2) {
if (o1 > o2) {
return o1;
} else {
return o2;
}
}
public static Matrix matrixToVector(Matrix matrix, boolean isRow) throws Exception {//将一个矩阵转成行向量
int x = matrix.getX();
int y = matrix.getY();

3
src/main/java/org/wlld/imageRecognition/Operation.java
Unescape View File

@ -203,7 +203,10 @@ public class Operation {//进行计算
Matrix yw = borderBody.getyW();
Matrix hw = borderBody.gethW();
Matrix ww = borderBody.getwW();
//将矩阵化为横向量
matrix = MatrixOperation.matrixToVector(matrix, true);
//最后加一层池化
matrix = MatrixOperation.getPoolVector(matrix);
//将参数矩阵的末尾填1
matrix = MatrixOperation.push(matrix, 1, true);
//锚点坐标及长宽预测值

2
src/main/java/org/wlld/imageRecognition/border/Border.java
Unescape View File

@ -67,6 +67,8 @@ public class Border {
double th = Math.log(ArithUtil.div(height, modelHeight));
//进行参数汇集 矩阵转化为行向量
matrix = MatrixOperation.matrixToVector(matrix, true);
//最后给一层池化层
matrix = MatrixOperation.getPoolVector(matrix);
//将参数矩阵的末尾填1
matrix = MatrixOperation.push(matrix, 1, true);
if (matrixX == null) {//如果是第一次直接赋值

58
src/test/java/org/wlld/HelloWorld.java
Unescape View File

@ -8,9 +8,11 @@ import org.wlld.imageRecognition.Operation;
import org.wlld.imageRecognition.Picture;
import org.wlld.imageRecognition.TempleConfig;
import org.wlld.imageRecognition.border.Frame;
import org.wlld.imageRecognition.border.FrameBody;
import org.wlld.nerveEntity.ModelParameter;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
/**
@ -28,34 +30,46 @@ public class HelloWorld {
public static void test() throws Exception {
Picture picture = new Picture();
TempleConfig templeConfig = new TempleConfig();
ModelParameter modelParameter = JSONObject.parseObject(ModelData.DATA2, ModelParameter.class);
templeConfig.setHavePosition(true);
Frame frame = new Frame();
frame.setWidth(3024);
frame.setHeight(4032);
frame.setLengthHeight(100);
frame.setLengthWidth(100);
templeConfig.setFrame(frame);
ModelParameter modelParameter = JSONObject.parseObject(ModelData.DATA, ModelParameter.class);
templeConfig.init(StudyPattern.Accuracy_Pattern, true, 3204, 4032, 1);
templeConfig.insertModel(modelParameter);
Operation operation = new Operation(templeConfig);
// for (int i = 1; i < 10; i++) {//faster rcnn神经网络学习
// System.out.println("study==" + i);
// //读取本地URL地址图片,并转化成矩阵
// Matrix right = picture.getImageMatrixByLocal("/Users/lidapeng/Desktop/myDocment/c/c" + i + ".png");
// Matrix wrong = picture.getImageMatrixByLocal("/Users/lidapeng/Desktop/myDocment/b/b" + i + ".png");
// //将图像矩阵和标注加入进行学习，Accuracy_Pattern 模式 进行第二次学习
// //第二次学习的时候，第三个参数必须是 true
// operation.learning(right, 1, true);
// operation.learning(wrong, 0, true);
// }
//templeConfig.clustering();//进行聚类
// ModelParameter modelParameter1 = templeConfig.getModel();
// String a = JSON.toJSONString(modelParameter1);
// System.out.println(a);
//测试集图片,进行识别测试
for (int i = 1; i < 300; i++) {//faster rcnn神经网络学习
System.out.println("study==" + i);
//读取本地URL地址图片,并转化成矩阵
Matrix right = picture.getImageMatrixByLocal("/Users/lidapeng/Desktop/myDocment/c/c" + i + ".png");
Matrix wrong = picture.getImageMatrixByLocal("/Users/lidapeng/Desktop/myDocment/b/b" + i + ".png");
//将图像矩阵和标注加入进行学习，Accuracy_Pattern 模式 进行第二次学习
//第二次学习的时候，第三个参数必须是 true
operation.learning(right, 1, true);
operation.learning(wrong, 0, true);
}
templeConfig.boxStudy();//边框回归
templeConfig.clustering();//进行聚类
ModelParameter modelParameter1 = templeConfig.getModel();
String a = JSON.toJSONString(modelParameter1);
System.out.println(a);
for (int j = 121; j < 140; j++) {
Matrix right = picture.getImageMatrixByLocal("/Users/lidapeng/Desktop/myDocment/c/c" + j + ".png");
Matrix wrong = picture.getImageMatrixByLocal("/Users/lidapeng/Desktop/myDocment/b/b" + j + ".png");
int rightId = operation.toSee(right);
int wrongId = operation.toSee(wrong);
System.out.println("该图是菜单：" + rightId);
System.out.println("该图是桌子:" + wrongId);
Map<Integer, List<FrameBody>> map = operation.lookWithPosition(right, j);
System.out.println("j===" + j);
}
//测试集图片,进行识别测试
// for (int j = 121; j < 140; j++) {
// Matrix right = picture.getImageMatrixByLocal("/Users/lidapeng/Desktop/myDocment/c/c" + j + ".png");
// Matrix wrong = picture.getImageMatrixByLocal("/Users/lidapeng/Desktop/myDocment/b/b" + j + ".png");
// int rightId = operation.toSee(right);
// int wrongId = operation.toSee(wrong);
// System.out.println("该图是菜单：" + rightId);
// System.out.println("该图是桌子:" + wrongId);
// }
}

23
src/test/java/org/wlld/MatrixTest.java
Unescape View File

@ -4,6 +4,8 @@ import org.wlld.MatrixTools.Matrix;
import org.wlld.MatrixTools.MatrixOperation;
import org.wlld.imageRecognition.border.BorderBody;
import java.util.Random;
/**
* @author lidapeng
* @description
@ -11,24 +13,15 @@ import org.wlld.imageRecognition.border.BorderBody;
*/
public class MatrixTest {
public static void main(String[] args) throws Exception {
double a = 3.33333;
double b = 3;
System.out.println(a / b);
//test3();
test4();
}
public static void test4() throws Exception {
BorderBody borderBody = new BorderBody();
Matrix xw = borderBody.getxW();
String a = "[1]#" +
"[3]#" +
"[5]#";
xw = new Matrix(3, 1, a);
borderBody.setxW(xw);
Matrix xt = borderBody.getxW();
xt = MatrixOperation.push(xt, 9, false);
Matrix xm = borderBody.getxW();
System.out.println(xm.getString());
Matrix matrix = new Matrix(1, 12);
String a = "[1,2,3,4,5,6,7,8,9,10,11,12]#";
matrix.setAll(a);
Matrix matrix1 = MatrixOperation.getPoolVector(matrix);
System.out.println(matrix1.getString());
}
public static void test3() throws Exception {

Write Preview
Loading…
Cancel
Save