public void GetDefaultName()
{
using (var model = SVM.Create())
{
Assert.Equal("opencv_ml_svm", model.GetDefaultName());
}
}
public void RunTest()
{
float[,] trainFeaturesData =
{
{ 0, 0 },
{ 0, 100 },
{ 100, 0 },
{ 100, 100 },
};
var trainFeatures = new Mat(4, 2, MatType.CV_32F, trainFeaturesData);
int[] trainLabelsData = { +1, -1, +1, -1 };
var trainLabels = new Mat(4, 1, MatType.CV_32S, trainLabelsData);
var model = SVM.Create();
model.Type = SVM.Types.CSvc;
model.KernelType = SVM.KernelTypes.Linear;
model.TermCriteria = new TermCriteria(CriteriaType.MaxIter, 100, 1e-6);
model.Train(trainFeatures, SampleTypes.RowSample, trainLabels);
float[] testFeatureData = { 90, 90 };
var testFeature = new Mat(1, 2, MatType.CV_32F, testFeatureData);
var detectedClass = (int)model.Predict(testFeature);
Assert.AreEqual(-1, detectedClass);
}
public void SaveLoadTest()
{
float[,] trainFeaturesData =
{
{ 0, 0 },
{ 0, 100 },
{ 100, 0 },
{ 100, 100 },
};
var trainFeatures = new Mat(4, 2, MatType.CV_32F, trainFeaturesData);
int[] trainLabelsData = { +1, -1, +1, -1 };
var trainLabels = new Mat(4, 1, MatType.CV_32S, trainLabelsData);
const string fileName = "svm.yml";
if (File.Exists(fileName))
{
File.Delete(fileName);
}
using (var model = SVM.Create())
{
model.Type = SVM.Types.CSvc;
model.KernelType = SVM.KernelTypes.Linear;
model.TermCriteria = new TermCriteria(CriteriaTypes.MaxIter, 100, 1e-6);
model.Train(trainFeatures, SampleTypes.RowSample, trainLabels);
model.Save(fileName);
}
Assert.True(File.Exists(fileName));
string content = File.ReadAllText(fileName);
//Console.WriteLine(content);
//Assert.DoesNotThrow
using (var model2 = SVM.Load(fileName))
{
GC.KeepAlive(model2);
}
using (var model2 = SVM.LoadFromString(content))
{
GC.KeepAlive(model2);
}
using (var fs = new FileStorage(fileName, FileStorage.Modes.Read))
using (var model2 = SVM.Create())
{
var node = fs["opencv_ml_svm"];
Assert.NotNull(node);
#pragma warning disable CS8604
model2.Read(node);
#pragma warning restore CS8604
}
}
//设置SVM对图片进行训练
public static bool Train(Mat samples, Mat responses)
{
svm = SVM.Create();
svm.Type = SVM.Types.CSvc;
svm.KernelType = SVM.KernelTypes.Linear;
svm.TermCriteria = new TermCriteria(CriteriaType.MaxIter, 10000, 1e-10);
IsReady = true;
return(svm.Train(samples, SampleTypes.RowSample, responses));
}
public static SVMClassifier Create(List <ImageFeatures> trainingData)
{
var svm = SVM.Create();
svm.Type = SVM.Types.CSvc;
svm.KernelType = SVM.KernelTypes.Rbf;
svm.TermCriteria = TermCriteria.Both(maxCount: 1000, epsilon: 0.000001);
svm.Gamma = 100.0;
svm.C = 1.0;
svm.Train(trainingData.ToTrainingMat(), SampleTypes.RowSample, trainingData.ToResponseMat());
return(new SVMClassifier(svm));
}
//分类文件与标签，参数为文件列表与标签结构体
public static bool Train(List <SVMFileInfo> file)
{
svm = SVM.Create();
svm.Type = SVM.Types.CSvc;
svm.KernelType = SVM.KernelTypes.Linear;
svm.TermCriteria = new TermCriteria(CriteriaType.MaxIter, 1000, 1e-10);
CharCategorySVM.IsReady = true;
Mat samples = new Mat(); //特征矩阵
Mat responses = new Mat(); //标签矩阵
foreach (SVMFileInfo fi in file)
{
List <string> fileNames = FileIO.OpenFile(fi.FilePath); //读到需要的文件路径列表
if (fileNames == null || fileNames.Count <= 0) //如果文件夹中一个图片都没有就返回
{
return(false);
}
//处理每个文件
foreach (string s in fileNames)
{
Mat matImg = new Mat(s, ImreadModes.Grayscale); //get image
Cv2.Threshold(matImg, matImg, 0, 255, ThresholdTypes.Otsu); //二值化
float[] feature = ComputeHogDescriptors(matImg); //提取图片HOG特征
samples.PushBack(TypeConvert.Float2Mat(feature)); //向特征矩阵中添加一行特征向量
responses.PushBack(TypeConvert.Int2Mat(fi.Label)); //向标签矩阵中添加一行标签
}
}
samples.ConvertTo(samples, MatType.CV_32FC1); //训练数据的格式必须是32位浮点型
if (svm.Train(samples, SampleTypes.RowSample, responses))
{
return(true);
}
return(false);
}
public static void Run()
{
Console.WriteLine("NuSVMDemo");
var class1 = DemoHelper.GenerateClass(0, 0.1, 0.1, 50);
var class2 = DemoHelper.GenerateClass(1, 0.8, 0.8, 50);
var trainData = class1.Concat(class2);
var trainer = SVM.Create(new NuSupportVectorClassification(0.1), new RbfKernel(0.5));
var model = trainer.Train(trainData.Select(p => Tuple.Create(p.ToArray(), p.Label)));
var x = new Point(0.9, 0.9).ToArray();
var resx = model.Predict(x);
Console.WriteLine(resx);
var y = new Point(0.0, 0.0).ToArray();
var resy = model.Predict(y);
Console.WriteLine(resy);
}
public static void Run()
{
Console.WriteLine("EpsSVRDemo");
var rnd = new Random();
var trainData = DemoHelper.Range(-10.0, 10.01, 0.1).Select(val => new { X = val, Y = DemoHelper.Sinc(val) + (rnd.NextDouble() - 0.5) / 4 });
var trainer = SVM.Create(new EpsilonSupportVectorRegression(1.0, 0.1), new RbfKernel(0.5));
var model = trainer.Train(trainData.Select(p => Tuple.Create(p.X.ToArray(), p.Y)));
foreach (var item in DemoHelper.Range(-1.0, 1.01, 0.1))
{
var x = item.ToArray();
var yPred = model.Predict(x);
var yReal = DemoHelper.Sinc(item);
Console.WriteLine("x: {0}", item);
Console.WriteLine("y_real: {0}", yReal);
Console.WriteLine("y_pred: {0}", yPred);
Console.WriteLine();
}
}
public static void Run()
{
Console.WriteLine("OneClassDemo");
var trainData = DemoHelper.GenerateClass(0, 0.5, 0.5, 100);
var trainer = SVM.Create(new OneClass(0.5), new RbfKernel(0.5));
var model = trainer.Train(trainData.Select(p => p.ToArray()));
var x = new Point(0.9, 0.9).ToArray();
var resx = model.Predict(x);
Console.WriteLine(resx);
var y = new Point(0.5, 0.5).ToArray();
var resy = model.Predict(y);
Console.WriteLine(resy);
var z = new Point(0.45, 0.45).ToArray();
var resz = model.Predict(z);
Console.WriteLine(resz);
}
public override void RunTest()
{
// Training data
var points = new Point2f[500];
var responses = new int[points.Length];
var rand = new Random();
for (int i = 0; i < responses.Length; i++)
{
float x = rand.Next(0, 300);
float y = rand.Next(0, 300);
points[i] = new Point2f(x, y);
responses[i] = (y > Function(x)) ? 1 : 2;
}
// Show training data and f(x)
using (Mat pointsPlot = Mat.Zeros(300, 300, MatType.CV_8UC3))
{
for (int i = 0; i < points.Length; i++)
{
int x = (int)points[i].X;
int y = (int)(300 - points[i].Y);
int res = responses[i];
Scalar color = (res == 1) ? Scalar.Red : Scalar.GreenYellow;
pointsPlot.Circle(x, y, 2, color, -1);
}
// f(x)
for (int x = 1; x < 300; x++)
{
int y1 = (int)(300 - Function(x - 1));
int y2 = (int)(300 - Function(x));
pointsPlot.Line(x - 1, y1, x, y2, Scalar.LightBlue, 1);
}
Window.ShowImages(pointsPlot);
}
// Train
var dataMat = new Mat(points.Length, 2, MatType.CV_32FC1, points);
var resMat = new Mat(responses.Length, 1, MatType.CV_32SC1, responses);
using var svm = SVM.Create();
// normalize data
dataMat /= 300.0;
// SVM parameters
svm.Type = SVM.Types.CSvc;
svm.KernelType = SVM.KernelTypes.Rbf;
svm.TermCriteria = TermCriteria.Both(1000, 0.000001);
svm.Degree = 100.0;
svm.Gamma = 100.0;
svm.Coef0 = 1.0;
svm.C = 1.0;
svm.Nu = 0.5;
svm.P = 0.1;
svm.Train(dataMat, SampleTypes.RowSample, resMat);
// Predict for each 300x300 pixel
using Mat retPlot = Mat.Zeros(300, 300, MatType.CV_8UC3);
for (int x = 0; x < 300; x++)
{
for (int y = 0; y < 300; y++)
{
float[] sample = { x / 300f, y / 300f };
var sampleMat = new Mat(1, 2, MatType.CV_32FC1, sample);
int ret = (int)svm.Predict(sampleMat);
var plotRect = new Rect(x, 300 - y, 1, 1);
if (ret == 1)
{
retPlot.Rectangle(plotRect, Scalar.Red);
}
else if (ret == 2)
{
retPlot.Rectangle(plotRect, Scalar.GreenYellow);
}
}
}
Window.ShowImages(retPlot);
}
static Plate_SVM()
{
svm = SVM.Create();
svm.Type = SVM.Types.CSvc;
svm.KernelType = SVM.KernelTypes.Linear;
}
/// <summary>
/// train model
/// </summary>
static void Train()
{
//load catalog
var paras = File.ReadAllLines("JPEGImages\\SVM_Train.txt");
var imgPath = new List <string>();
var imgCatg = new List <int>();
var nImgCnt = paras.Length;
foreach (var i in paras)
{
var temp = i.Split(':');
imgPath.Add(temp[1]);
imgCatg.Add(Convert.ToInt32(temp[0]));
}
//create HOG
var hog = new HOGDescriptor(new Size(64, 64), new Size(16, 16), new Size(8, 8), new Size(8, 8));
Mat data = Mat.Zeros(nImgCnt, hog.GetDescriptorSize(), MatType.CV_32FC1);
Mat res = Mat.Zeros(nImgCnt, 1, MatType.CV_32SC1);
for (var z = 0; z < nImgCnt; z++)
{
//load img
Mat src = Cv2.ImRead(imgPath[z], ImreadModes.GrayScale);
Console.WriteLine($"Processing: {Path.GetFileNameWithoutExtension(imgPath[z])}");
//resize to 64*64
Cv2.Resize(src, src, new Size(64, 64));
//threshold
src = src.Threshold(200, 255, ThresholdTypes.Binary);
//center image
MoveToCenter(src);
//computer descriptors
var descriptors = hog.Compute(src, new Size(1, 1), new Size(0, 0));
for (var i = 0; i < descriptors.Length; i++)
{
data.Set(z, i, descriptors[i]);
}
res.Set(z, 0, imgCatg[z]);
src.Release();
}
Console.WriteLine("Start training");
//create svm
var svm = SVM.Create();
svm.TermCriteria = new TermCriteria(CriteriaType.Eps, 1000, float.Epsilon);
svm.Type = SVM.Types.CSvc;
svm.KernelType = SVM.KernelTypes.Rbf;
svm.Degree = 10;
svm.Gamma = 0.09;
svm.Coef0 = 1;
svm.C = 10;
svm.Nu = 0.5;
svm.P = 1;
//training
svm.Train(data, SampleTypes.RowSample, res);
//save result
svm.Save("SVM_RESULT.xml");
}