public override int ProcessImages(List <ImageData> images, out double[][] inputs, out int[] outputs)
{
outputs = new int[images.Count];
inputs = new double[images.Count][];
var imageData = new List <Bitmap>();
foreach (var image in images)
{
using (var stream = new MemoryStream(image.Image))
{
imageData.Add(new Bitmap(stream));
}
}
var converter = new CornerFeaturesDetector(_detector);
_bagOfVisualWords = new BagOfVisualWords <CornerFeaturePoint>(converter, NumberOfWords);
_bagOfVisualWords.Compute(imageData.ToArray());
for (var i = 0; i < images.Count; i++)
{
inputs[i] = _bagOfVisualWords.GetFeatureVector(imageData[i]);
outputs[i] = Utilities.GetIntFromClass(images[i].Type);
}
return(images.Count);
}
public void BagOfVisualWordsConstructorTest()
{
BagOfVisualWords bow = new BagOfVisualWords(10);
var points = bow.Compute(images, 1e-3);
Assert.AreEqual(10, bow.NumberOfWords);
Assert.AreEqual(6, points.Length);
Assert.AreEqual(406, points[0].Count);
Assert.AreEqual(727, points[1].Count);
Assert.AreEqual(549, points[2].Count);
Assert.AreEqual(458, points[3].Count);
Assert.AreEqual(723, points[4].Count);
Assert.AreEqual(1263, points[5].Count);
Assert.AreEqual(388.043776954555, points[0][0].X);
Assert.AreEqual(105.99327164889745, points[0][0].Y);
Assert.AreEqual(335.64548481033881, points[3][7].X);
Assert.AreEqual(152.14505651866821, points[2][3].Y);
Assert.AreEqual(573.691355494602, points[2][52].X);
Assert.AreEqual(153.6650841848263, points[1][11].Y);
Assert.AreEqual(573.03087205188058, points[0][42].X);
Assert.AreEqual(374.27580307739436, points[4][125].Y);
}
/// <summary>
/// Initialise a Visual Bag of Words Model
/// </summary>
/// <returns></returns>
public static BagOfVisualWords <Accord.IFeatureDescriptor <double[]>, double[], Accord.MachineLearning.KMeans, Accord.Imaging.FastRetinaKeypointDetector> CreateBowModel()
{
if (Image.targetBitmap == null)
{
System.Drawing.Image imageTarget = System.Drawing.Image.FromFile(@"C:\Users\antmi\Pictures\bark1.jpg");
Size size = new Size(200, 200);
Image.targetBitmap = new Bitmap(imageTarget, size);
}
var bow = BagOfVisualWords.Create(new FastRetinaKeypointDetector(), numberOfWords: 10);
Bitmap[] images = new Bitmap[1] {
Image.targetBitmap
};
bow.Learn(images);
if (Image.targetBitmapScaled == null)
{
System.Drawing.Image imageTarget = System.Drawing.Image.FromFile(@"C:\Users\antmi\Pictures\bark1.jpg");
Size size = new Size(100, 100);
Image.targetBitmapScaled = new Bitmap(imageTarget, size);
Image.scaledFeatureCount = bow.Transform(new Bitmap[1] {
Image.targetBitmapScaled
})[0].Sum();
}
return(bow);
}
public static void createBoW()
{
string[] files = Directory.GetFiles ("testset");
Dictionary<string, Bitmap> testImages = new Dictionary<string, Bitmap>();
for (int i = 0; i < files.Length; i++) {
string name = Path.GetFileNameWithoutExtension (files [i]);
testImages.Add(name, (Bitmap)Bitmap.FromFile(files[i]));
}
int numberOfWords = 6; // number of cluster centers: typically >>100
// Create a Binary-Split clustering algorithm
BinarySplit binarySplit = new BinarySplit(numberOfWords);
// Create bag-of-words (BoW) with the given algorithm
BagOfVisualWords surfBow = new BagOfVisualWords(binarySplit);
// Compute the BoW codebook using training images only
Bitmap[] bmps = new Bitmap[testImages.Count];
testImages.Values.CopyTo(bmps, 0);
surfBow.Compute(bmps);
surfBow.Save ("bagOfWords");
}
public void SerializeTest()
{
var images = BagOfVisualWordsTest.images.DeepClone();
Accord.Math.Tools.SetupGenerator(0);
BagOfVisualWords bow = new BagOfVisualWords(10);
bow.Compute(images);
double[][] expected = new double[images.Length][];
for (int i = 0; i < expected.Length; i++)
{
expected[i] = bow.GetFeatureVector(images[i]);
}
MemoryStream stream = new MemoryStream();
BinaryFormatter fmt = new BinaryFormatter();
fmt.Serialize(stream, bow);
stream.Seek(0, SeekOrigin.Begin);
bow = (BagOfVisualWords)fmt.Deserialize(stream);
double[][] actual = new double[expected.Length][];
for (int i = 0; i < actual.Length; i++)
{
actual[i] = bow.GetFeatureVector(images[i]);
}
Assert.IsTrue(expected.IsEqual(actual));
}
public void BagOfVisualWordsConstructorTest()
{
BagOfVisualWords bow = new BagOfVisualWords(10);
var points = bow.Compute(images, 1e-3);
Assert.AreEqual(10, bow.NumberOfWords);
Assert.AreEqual(6, points.Length);
Assert.AreEqual(406, points[0].Count);
Assert.AreEqual(727, points[1].Count);
Assert.AreEqual(549, points[2].Count);
Assert.AreEqual(458, points[3].Count);
Assert.AreEqual(723, points[4].Count);
Assert.AreEqual(1263, points[5].Count);
Assert.AreEqual(388.043776954555, points[0][0].X);
Assert.AreEqual(105.99327164889745, points[0][0].Y);
Assert.AreEqual(335.64548481033881, points[3][7].X);
Assert.AreEqual(152.14505651866821, points[2][3].Y);
Assert.AreEqual(573.691355494602, points[2][52].X);
Assert.AreEqual(153.6650841848263, points[1][11].Y);
Assert.AreEqual(573.03087205188058, points[0][42].X);
Assert.AreEqual(374.27580307739436, points[4][125].Y);
}
private IBagOfWords <Bitmap> CreateBow()
{
var binarySplit = new BinarySplit(10);
var surfBow = BagOfVisualWords.Create(10);
return(surfBow.Learn(Images.ToArray()));
}
private static void CreateBoW()
{
var numberOfWords = 36;
foreach (var file in Directory.EnumerateFiles(@"C:\Temp\TLLCamerasTestData\37_Training", "*.jpg"))
{
var trainingImage = (Bitmap)Bitmap.FromFile(file);
trainingImages.Add(file, trainingImage);
}
foreach (var file in Directory.EnumerateFiles(@"C:\Temp\TLLCamerasTestData\37_Testing", "*.jpg"))
{
var testImage = (Bitmap)Bitmap.FromFile(file);
testingImages.Add(file, testImage);
}
// We will use SURF, so we can use a standard clustering
// algorithm that is based on Euclidean distances. A good
// algorithm for clustering codewords is the Binary Split
// variant of the K-Means algorithm.
// Create a Binary-Split clustering algorithm
BinarySplit binarySplit = new BinarySplit(numberOfWords);
// Create bag-of-words (BoW) with the given algorithm
BagOfVisualWords surfBow = new BagOfVisualWords(binarySplit);
// Compute the BoW codebook using training images only
IBagOfWords <Bitmap> bow = surfBow.Learn(trainingImages.Values.ToArray());
// now that we've created the bow we need to use it to create a representation of each training and test image
foreach (var trainingImage in trainingImages.Keys)
{
var asBitmap = trainingImages[trainingImage] as Bitmap;
var featureVector = (bow as ITransform <Bitmap, double[]>).Transform(asBitmap);
var featureString = featureVector.ToString(DefaultArrayFormatProvider.InvariantCulture);
trainingFeatures.Add(trainingImage, featureVector);
}
foreach (var testingImage in testingImages.Keys)
{
var asBitmap = testingImages[testingImage] as Bitmap;
var featureVector = (bow as ITransform <Bitmap, double[]>).Transform(asBitmap);
var featureString = featureVector.ToString(DefaultArrayFormatProvider.InvariantCulture);
testingFeatures.Add(testingImage, featureVector);
}
}
private void bowsToolStripMenuItem_Click_1(object sender, EventArgs e)
{
var bow = BagOfVisualWords.Create(numberOfWords: bowSize);
var images = trainData.GetBitmaps(mask);
bow.Learn(images);
Accord.IO.Serializer.Save(bow, dataPath + String.Format(@"\train-{0}.bow", bowSize));
logger.logStr("Done " + bowSize);
Application.DoEvents();
}
public void CalculateSampleFeatures(BagOfVisualWords bow)
{
List<double[]> features = new List<double[]>();
foreach(var img in Images)
{
var vector = bow.GetFeatureVector(img.BitmapImage);
features.Add(vector);
}
SampleFeatures = features.ToArray();
}
/// <summary>
/// This methods computes the Bag-of-Visual-Words with the training images.
/// </summary>
///
private void btnBagOfWords_Click(object sender, EventArgs e)
{
int numberOfWords = (int)numWords.Value;
// Create a Binary-Split clustering algorithm
BinarySplit binarySplit = new BinarySplit(numberOfWords);
// Create bag-of-words (BoW) with the given algorithm
BagOfVisualWords bow = new BagOfVisualWords(binarySplit);
if (cbExtended.Checked)
{
bow.Detector.ComputeDescriptors = SpeededUpRobustFeatureDescriptorType.Extended;
}
Stopwatch sw1 = Stopwatch.StartNew();
// Compute the BoW codebook using training images only
var points = bow.Compute(originalTrainImages.Values.ToArray());
sw1.Stop();
Stopwatch sw2 = Stopwatch.StartNew();
// Extract features for all images
foreach (ListViewItem item in listView1.Items)
{
// Get item image
Bitmap image = originalImages[item.ImageKey] as Bitmap;
// Process image
double[] featureVector = bow.GetFeatureVector(image);
string featureString = featureVector.ToString(DefaultArrayFormatProvider.InvariantCulture);
if (item.SubItems.Count == 2)
{
item.SubItems[1].Text = featureString;
}
else
{
item.SubItems.Add(featureString);
}
int classLabel = (item.Tag as Tuple <double[], int>).Item2;
item.Tag = Tuple.Create(featureVector, classLabel);
}
sw2.Stop();
lbStatus.Text = "BoW constructed in " + sw1.Elapsed + "s. Features extracted in " + sw2.Elapsed + "s.";
btnSampleRunAnalysis.Enabled = true;
}
/// <summary>
/// Computes the Bag of Words model.
/// </summary>
/// <typeparam name="TPoint">
/// The <see cref="Accord.Imaging.IFeaturePoint{TFeature}"/> type to be used with this class,
/// such as <see cref="Accord.Imaging.SpeededUpRobustFeaturePoint"/>.
/// </typeparam>
/// <typeparam name="TFeature">
/// The feature type of the <typeparamref name="TPoint"/>, such
/// as <see cref="T:double[]"/>.
/// </typeparam>
/// <param name="bow">Bag of Visual Words.</param>
/// <param name="images">The set of images to initialize the model.</param>
/// <param name="threshold">Convergence rate for the k-means algorithm. Default is 1e-5.</param>
/// <returns>The list of feature points detected in all images.</returns>
public static List <TPoint>[] Compute <TPoint, TFeature>(this BagOfVisualWords <TPoint, TFeature> bow,
Image <Gray, byte>[] images, double threshold = 1e-5)
where TPoint : IFeatureDescriptor <TFeature>
{
var featurePoints = bow.Compute
(
images.Select(x => x.ToBitmap(copyAlways: false, failIfCannotCast: true)).ToArray(),
threshold
);
return(featurePoints);
}
public bool TrainSceneFeature(SceneFeatureData scenefeatureData)
{
//Create bow
Bitmap mask = Utils.CreateMaskBitmap(new Size(1280, 720), new Rectangle[] { scenefeatureData.feature.area });
var bow = BagOfVisualWords.Create(numberOfWords: scenefeatureData.feature.bowSize);
var images = scenefeatureData.trainData.GetBitmaps(mask);
bow.Learn(images);
Accord.IO.Serializer.Save(bow, path + @"\" + scenefeatureData.feature.name + String.Format(@"\train-{0}.bow", scenefeatureData.feature.bowSize));
bow.Show();
return(Train(bow, scenefeatureData));
}
public void GetFeatureVectorTest()
{
var images = GetImages();
Accord.Math.Random.Generator.Seed = 0;
Bitmap image = new Bitmap(images[0]);
// The Bag-of-Visual-Words model converts arbitrary-size images
// into fixed-length feature vectors. In this example, we will
// be setting the codebook size to 10. This means all generated
// feature vectors will have the same length of 10.
// Create a new Bag-of-Visual-Words (BoW) model
BagOfVisualWords bow = new BagOfVisualWords(10);
bow.ParallelOptions.MaxDegreeOfParallelism = 1;
// Compute the model using
// a set of training images
bow.Compute(images);
// After this point, we will be able to translate
// images into double[] feature vectors using
double[] feature = bow.GetFeatureVector(image);
Assert.AreEqual(10, feature.Length);
double[][] expected =
{
new double[] { 47, 44, 42, 4, 23, 22, 28, 53, 50, 96 },
new double[] { 26, 91, 71, 49, 99, 70, 59, 28, 155, 79 },
new double[] { 71, 34, 51, 33, 53, 25, 44, 64, 32, 145 }
};
double[][] actual = new double[expected.Length][];
for (int i = 0; i < actual.Length; i++)
{
actual[i] = bow.GetFeatureVector(images[i]);
}
//string str = actual.ToString(CSharpJaggedMatrixFormatProvider.InvariantCulture);
//Assert.IsNullOrEmpty(str);
for (int i = 0; i < actual.Length; i++)
{
for (int j = 0; j < actual[i].Length; j++)
{
Assert.IsTrue(expected[i].Contains(actual[i][j]));
}
}
}
public static double[][] GetSURFFeatures(Bitmap[] images, int maxFeatures, string modelSaveLocation = null)
{
Accord.Math.Random.Generator.Seed = 0;
var bow = BagOfVisualWords.Create(numberOfWords: maxFeatures);
bow.Learn(images);
if (!string.IsNullOrEmpty(modelSaveLocation))
{
Serializer.Save(obj: bow, path: modelSaveLocation);
}
double[][] features = bow.Transform(images);
return(features);
}
static void ImageFeatures()
{
Dictionary <string, Bitmap> testImages = new Dictionary <string, Bitmap>();
testImages.Add("img_acropolis", (Bitmap)Bitmap.FromFile("test_imgs/acropolis_athens.jpg"));
testImages.Add("img_cathedral", (Bitmap)Bitmap.FromFile("test_imgs/amiens_cathedral.jpg"));
testImages.Add("img_bigben", (Bitmap)Bitmap.FromFile("test_imgs/big_ben.jpg"));
int numberOfWords = 6; // number of cluster centers: typically >>100
// Create a Binary-Split clustering algorithm
BinarySplit binarySplit = new BinarySplit(numberOfWords);
IBagOfWords <Bitmap> bow;
// Create bag-of-words ( BoW ) with the given algorithm
BagOfVisualWords surfBow = new BagOfVisualWords(binarySplit);
// Compute the BoW codebook using training images only
Bitmap[] bmps = new Bitmap[testImages.Count];
testImages.Values.CopyTo(bmps, 0);
surfBow.Compute(bmps);
bow = surfBow;
// this model needs to be saved once it is calculated: only compute it once to calculate features
// from the collection as well as for new queries.
// THE SAME TRAINED MODEL MUST BE USED TO GET THE SAME FEATURES!!!
Dictionary <string, double[]> testImageFeatures = new Dictionary <string, double[]>();
// Extract features for all images
foreach (string imagename in testImages.Keys)
{
double[] featureVector = bow.GetFeatureVector(testImages[imagename]);
testImageFeatures.Add(imagename, featureVector);
Console.Out.WriteLine(imagename + " features: " + featureVector.ToString(DefaultArrayFormatProvider.InvariantCulture));
}
// Calculate Image Similarities
string[] imagenames = new string[testImageFeatures.Keys.Count];
testImageFeatures.Keys.CopyTo(imagenames, 0);
for (int i = 0; i < imagenames.Length; i++)
{
for (int j = i + 1; j < imagenames.Length; j++)
{
double dist = Distance.Cosine(testImageFeatures[imagenames[i]], testImageFeatures[imagenames[j]]);
Console.Out.WriteLine(imagenames[i] + " <-> " + imagenames[j] + " distance: " + dist.ToString());
}
}
}
public void GetFeatureVectorTest()
{
var images = GetImages();
Accord.Math.Random.Generator.Seed = 0;
// The Bag-of-Visual-Words model converts arbitrary-size images
// into fixed-length feature vectors. In this example, we will
// be setting the codebook size to 10. This means all generated
// feature vectors will have the same length of 10.
// Create a new Bag-of-Visual-Words (BoW) model
BagOfVisualWords bow = new BagOfVisualWords(10);
bow.ParallelOptions.MaxDegreeOfParallelism = 1;
// Compute the model using
// a set of training images
bow.Compute(images);
// After this point, we will be able to translate
// images into double[] feature vectors using
double[] feature = bow.GetFeatureVector(images[0]);
Assert.AreEqual(10, feature.Length);
double[][] expected = new double[][]
{
new double[] { 4, 28, 24, 68, 51, 97, 60, 35, 18, 24 },
new double[] { 53, 111, 89, 70, 24, 80, 130, 46, 50, 74 },
new double[] { 31, 29, 57, 102, 63, 142, 40, 18, 37, 33 }
};
double[][] actual = new double[expected.Length][];
for (int i = 0; i < actual.Length; i++)
{
actual[i] = bow.GetFeatureVector(images[i]);
}
string str = actual.ToCSharp();
for (int i = 0; i < actual.Length; i++)
{
for (int j = 0; j < actual[i].Length; j++)
{
Assert.IsTrue(expected[i].Contains(actual[i][j]));
}
}
}
public void LargeTest()
{
// Requires data from the National Data Science bowl
// https://github.com/accord-net/framework/issues/58
var trainingDirectory = @"C:\Users\CésarRoberto\Downloads\train\train\";
var images = LabeledImages.FromDirectory(trainingDirectory).ToArray();
var binarySplit = new BinarySplit(32);
var bow = new BagOfVisualWords(binarySplit);
bow.Compute(images.Select(i => i.Item).Take(50).ToArray());
}
private void loadBow(object sender, RoutedEventArgs e)
{
imageInfo.Text = "Loading BoW";
Microsoft.Win32.OpenFileDialog op = new Microsoft.Win32.OpenFileDialog();
op.Title = "Select a BoW file.";
if (op.ShowDialog() == true)
{
bow = Accord.IO.Serializer.Load <BagOfVisualWords>(op.FileName);
}
else
{
MessageBox.Show("Something went wrong while loadin BoW!");
}
imageInfo.Text = "Done";
}
/// <summary>
/// Computes the Bag of Words model.
/// </summary>
/// <typeparam name="TPoint">
/// The <see cref="Accord.Imaging.IFeaturePoint{TFeature}"/> type to be used with this class,
/// such as <see cref="Accord.Imaging.SpeededUpRobustFeaturePoint"/>.
/// </typeparam>
/// <typeparam name="TFeature">
/// The feature type of the <typeparamref name="TPoint"/>, such
/// as <see cref="T:double[]"/>.
/// </typeparam>
/// <param name="bow">Bag of Visual Words.</param>
/// <param name="images">The set of images to initialize the model.</param>
/// <param name="threshold">Convergence rate for the k-means algorithm. Default is 1e-5.</param>
/// <returns>The list of feature points detected in all images.</returns>
public static List <TPoint>[] Compute <TPoint, TFeature>(this BagOfVisualWords <TPoint, TFeature> bow,
Gray <byte>[][,] images, double threshold = 1e-5)
where TPoint : IFeatureDescriptor <TFeature>
{
var uImages = images.Select(x => x.Lock());
var featurePoints = bow.Compute
(
uImages.Select(x => x.AsBitmap()).ToArray(),
threshold
);
uImages.ForEach(x => x.Dispose());
return(featurePoints);
}
public void GetFeatureVectorTest()
{
Accord.Math.Tools.SetupGenerator(0);
Bitmap image = (Bitmap)images[0].Clone();
// The Bag-of-Visual-Words model converts arbitrary-size images
// into fixed-length feature vectors. In this example, we will
// be setting the codebook size to 10. This means all generated
// feature vectors will have the same length of 10.
// Create a new Bag-of-Visual-Words (BoW) model
BagOfVisualWords bow = new BagOfVisualWords(10);
// Compute the model using
// a set of training images
bow.Compute(images);
// After this point, we will be able to translate
// images into double[] feature vectors using
double[] feature = bow.GetFeatureVector(image);
Assert.AreEqual(10, feature.Length);
double[][] expected =
{
new double[] { 43, 39, 27, 21, 34, 30, 98, 51, 61, 5 },
new double[] { 69, 100, 61, 41, 114, 125, 85, 27, 34, 71 },
new double[] { 52, 46, 44, 19, 20, 51, 143, 73, 69, 35 }
};
double[][] actual = new double[expected.Length][];
for (int i = 0; i < actual.Length; i++)
{
actual[i] = bow.GetFeatureVector(images[i]);
}
// string str = actual.ToString(CSharpJaggedMatrixFormatProvider.InvariantCulture);
for (int i = 0; i < actual.Length; i++)
{
for (int j = 0; j < actual[i].Length; j++)
{
Assert.IsTrue(expected[i].Contains(actual[i][j]));
}
}
}
public void GetFeatureVectorTest()
{
Accord.Math.Tools.SetupGenerator(0);
Bitmap image = images[0];
// The Bag-of-Visual-Words model converts arbitrary-size images
// into fixed-length feature vectors. In this example, we will
// be setting the codebook size to 10. This means all generated
// feature vectors will have the same length of 10.
// Create a new Bag-of-Visual-Words (BoW) model
BagOfVisualWords bow = new BagOfVisualWords(10);
// Compute the model using
// a set of training images
bow.Compute(images);
// After this point, we will be able to translate
// images into double[] feature vectors using
double[] feature = bow.GetFeatureVector(image);
Assert.AreEqual(10, feature.Length);
double[][] expected =
{
new double[] { 102, 58, 42, 24, 47, 53, 29, 26, 3, 22 },
new double[] { 90, 135, 71, 101, 26, 28, 105, 61, 62, 48 },
new double[] { 138, 36, 55, 56, 71, 61, 30, 47, 33, 22 }
};
double[][] actual = new double[expected.Length][];
for (int i = 0; i < actual.Length; i++)
{
actual[i] = bow.GetFeatureVector(images[i]);
}
// string str = actual.ToString(CSharpJaggedMatrixFormatProvider.InvariantCulture);
for (int i = 0; i < actual.Length; i++)
{
for (int j = 0; j < actual[i].Length; j++)
{
Assert.IsTrue(expected[i].Contains(actual[i][j]));
}
}
}
/// <summary>
/// Method use for deserialize bcf from file
/// </summary>
/// <returns>deserialized object</returns>
internal static BagOfVisualWords ReadBinary()
{
//Create binary formatter
BinaryFormatter formatter = new BinaryFormatter();
//Open stream for read
FileStream fs = new FileStream(Path.Combine(Application.StartupPath, "Resources/" +
"bcf.dat"), FileMode.OpenOrCreate);
//Deserialize data from strim and convert it in bcf
BagOfVisualWords freak = (BagOfVisualWords)formatter.Deserialize(fs);
//Close stream
fs.Close();
return(freak);
}
public void BagOfVisualWordsConstructorTest()
{
Accord.Math.Random.Generator.Seed = 0;
var images = GetImages();
BagOfVisualWords bow = new BagOfVisualWords(10);
bow.ParallelOptions.MaxDegreeOfParallelism = 1;
#pragma warning disable 612, 618
var points = bow.Compute(images, 1e-3);
#pragma warning restore 612, 618
Assert.AreEqual(10, bow.NumberOfWords);
Assert.AreEqual(6, points.Length);
Assert.AreEqual(409, points[0].Count);
Assert.AreEqual(727, points[1].Count);
Assert.AreEqual(552, points[2].Count);
Assert.AreEqual(460, points[3].Count);
Assert.AreEqual(719, points[4].Count);
Assert.AreEqual(1265, points[5].Count);
double tol = 1e-5;
Assert.AreEqual(388.04225639880224, points[0][0].X, tol);
Assert.AreEqual(105.9954439039073, points[0][0].Y, tol);
Assert.AreEqual(335.62395561144433, points[3][7].X, tol);
Assert.AreEqual(152.14505651866821, points[2][3].Y, tol);
Assert.AreEqual(573.691355494602, points[2][52].X, tol);
Assert.AreEqual(153.6650841848263, points[1][11].Y, tol);
Assert.AreEqual(289.54728415724327, points[0][42].X, tol);
Assert.AreEqual(373.99402540151056, points[4][125].Y, tol);
foreach (var point in points)
{
foreach (var p in point)
{
Assert.IsFalse(double.IsNaN(p.X));
Assert.IsFalse(double.IsNaN(p.Y));
}
}
}
public void BagOfVisualWordsConstructorTest3()
{
Accord.Math.Random.Generator.Seed = 0;
var images = GetImages();
MoravecCornersDetector moravec = new MoravecCornersDetector();
CornerFeaturesDetector detector = new CornerFeaturesDetector(moravec);
var bow = new BagOfVisualWords <CornerFeaturePoint>(detector, numberOfWords: 10);
bow.ParallelOptions.MaxDegreeOfParallelism = 1;
#pragma warning disable 612, 618
var points = bow.Compute(images, 1e-3);
#pragma warning restore 612, 618
double[] vector = bow.GetFeatureVector(images[0]);
Assert.AreEqual(10, bow.NumberOfWords);
Assert.AreEqual(6, points.Length);
Assert.AreEqual(10, vector.Length);
Assert.AreEqual(2800, points[0].Count);
Assert.AreEqual(4532, points[1].Count);
Assert.AreEqual(2282, points[2].Count);
Assert.AreEqual(1173, points[3].Count);
Assert.AreEqual(4860, points[4].Count);
Assert.AreEqual(5730, points[5].Count);
Assert.AreEqual(596, points[0][0].Descriptor[0]);
Assert.AreEqual(51, points[0][0].Descriptor[1]);
Assert.AreEqual(points[0][0].Descriptor[0], points[0][0].X);
Assert.AreEqual(points[0][0].Descriptor[1], points[0][0].Y);
Assert.AreEqual(461, points[3][7].Descriptor[0]);
Assert.AreEqual(8, points[2][3].Descriptor[1]);
Assert.AreEqual(991, points[2][52].Descriptor[0]);
Assert.AreEqual(82, points[1][11].Descriptor[1]);
Assert.AreEqual(430, points[0][42].Descriptor[0]);
Assert.AreEqual(135, points[4][125].Descriptor[1]);
}
public void GetFeatureVectorTest()
{
Accord.Math.Tools.SetupGenerator(0);
Bitmap image = images[0];
// The Bag-of-Visual-Words model converts arbitrary-size images
// into fixed-length feature vectors. In this example, we will
// be setting the codebook size to 10. This means all generated
// feature vectors will have the same length of 10.
// Create a new Bag-of-Visual-Words (BoW) model
BagOfVisualWords bow = new BagOfVisualWords(10);
// Compute the model using
// a set of training images
bow.Compute(images);
// After this point, we will be able to translate
// images into double[] feature vectors using
double[] feature = bow.GetFeatureVector(image);
Assert.AreEqual(10, feature.Length);
double[][] expected =
{
new double[] { 26, 58, 102, 24, 53, 22, 47, 29, 42, 3 },
new double[] { 61, 135, 90, 101, 28, 48, 26, 105, 71, 62 },
new double[] { 47, 36, 138, 56, 61, 22, 71, 30, 55, 33 },
};
double[][] actual = new double[expected.Length][];
for (int i = 0; i < actual.Length; i++)
{
actual[i] = bow.GetFeatureVector(images[i]);
}
Assert.IsTrue(expected.IsEqual(actual));
}
/// <summary>
/// Main constructor for windows form
/// </summary>
public MainWindow()
{
InitializeComponent();
//Add properties for openFileDialog
openFileDialog.DefaultExt = ".tif";
openFileDialog.Title = "Open Image";
openFileDialog.Filter = "JPEG files (*.jpg)| *.jpg|TIFF files (*.tif)| *.tif|" +
" All files | *.*";
//Close button Yes and No, which user click for assert recognition
labelCorrect.Visible = false;
buttonCorrectYes.Visible = false;
buttonCorrectYes.Visible = false;
this.BackColor = Color.White;
//Read ready SVM and bagOfContourFragments
multiSVM = BinarySave.ReadBinary(true);
bagOfContourFragments = BinarySave.ReadBinary();
}
/* public static Bitmap[] GetBitmaps(this ImageFileData[] _self)
* {
* List<Bitmap> ret = new List<Bitmap>();
* foreach (var a in _self)
* {
* ret.Add(a.image);
* }
* return ret.ToArray();
* }*/
public static void Show(this BagOfVisualWords bow)
{
// We can also check some statistics about the dataset:
int numberOfImages = bow.Statistics.TotalNumberOfInstances; // 6
// Statistics about all the descriptors that have been extracted:
int totalDescriptors = bow.Statistics.TotalNumberOfDescriptors; // 4132
double totalMean = bow.Statistics.TotalNumberOfDescriptorsPerInstance.Mean; // 688.66666666666663
double totalVar = bow.Statistics.TotalNumberOfDescriptorsPerInstance.Variance; // 96745.866666666669
IntRange totalRange = bow.Statistics.TotalNumberOfDescriptorsPerInstanceRange; // [409, 1265]
Console.WriteLine(String.Format("{0} {1} {2} {3}", totalDescriptors, totalMean, totalVar, totalRange.ToString()));
// Statistics only about the descriptors that have been actually used:
int takenDescriptors = bow.Statistics.NumberOfDescriptorsTaken; // 4132
double takenMean = bow.Statistics.NumberOfDescriptorsTakenPerInstance.Mean; // 688.66666666666663
double takenVar = bow.Statistics.NumberOfDescriptorsTakenPerInstance.Variance; // 96745.866666666669
IntRange takenRange = bow.Statistics.NumberOfDescriptorsTakenPerInstanceRange; // [409, 1265]
Console.WriteLine(String.Format("{0} {1} {2} {3}", takenDescriptors, takenMean, takenVar, takenRange.ToString()));
}
public void SerializeTest2()
{
var images = GetImages();
Accord.Math.Random.Generator.Seed = 0;
FastCornersDetector fast = new FastCornersDetector();
FastRetinaKeypointDetector freak = new FastRetinaKeypointDetector(fast);
var kmodes = new KModes <byte>(5, new Hamming());
kmodes.ParallelOptions.MaxDegreeOfParallelism = 1;
var bow = new BagOfVisualWords <FastRetinaKeypoint, byte[]>(freak, kmodes);
bow.Compute(images);
double[][] expected = new double[images.Length][];
for (int i = 0; i < expected.Length; i++)
{
expected[i] = bow.GetFeatureVector(images[i]);
}
MemoryStream stream = new MemoryStream();
BinaryFormatter fmt = new BinaryFormatter();
fmt.Serialize(stream, bow);
stream.Seek(0, SeekOrigin.Begin);
bow = (BagOfVisualWords <FastRetinaKeypoint, byte[]>)fmt.Deserialize(stream);
double[][] actual = new double[expected.Length][];
for (int i = 0; i < actual.Length; i++)
{
actual[i] = bow.GetFeatureVector(images[i]);
}
Assert.IsTrue(expected.IsEqual(actual));
}
public void BagOfVisualWordsConstructorTest()
{
BagOfVisualWords bow = new BagOfVisualWords(10);
var points = bow.Compute(images, 1e-3);
Assert.AreEqual(10, bow.NumberOfWords);
Assert.AreEqual(6, points.Length);
Assert.AreEqual(409, points[0].Count);
Assert.AreEqual(727, points[1].Count);
Assert.AreEqual(552, points[2].Count);
Assert.AreEqual(460, points[3].Count);
Assert.AreEqual(719, points[4].Count);
Assert.AreEqual(1265, points[5].Count);
double tol = 1e-7;
Assert.AreEqual(388.04225639880224, points[0][0].X, tol);
Assert.AreEqual(105.9954439039073, points[0][0].Y, tol);
Assert.AreEqual(335.62395561144433, points[3][7].X, tol);
Assert.AreEqual(152.14505651866821, points[2][3].Y, tol);
Assert.AreEqual(573.691355494602, points[2][52].X, tol);
Assert.AreEqual(153.6650841848263, points[1][11].Y, tol);
Assert.AreEqual(289.54728415724327, points[0][42].X, tol);
Assert.AreEqual(373.99402540151056, points[4][125].Y, tol);
foreach (var point in points)
{
foreach (var p in point)
{
Assert.IsFalse(double.IsNaN(p.X));
Assert.IsFalse(double.IsNaN(p.Y));
}
}
}
public void SerializeTest2()
{
var images = BagOfVisualWordsTest.images.DeepClone();
Accord.Math.Tools.SetupGenerator(0);
FastCornersDetector fast = new FastCornersDetector();
FastRetinaKeypointDetector freak = new FastRetinaKeypointDetector(fast);
var kmodes = new KModes <byte>(5, Distance.BitwiseHamming);
var bow = new BagOfVisualWords <FastRetinaKeypoint, byte[]>(freak, kmodes);
bow.Compute(images);
double[][] expected = new double[images.Length][];
for (int i = 0; i < expected.Length; i++)
{
expected[i] = bow.GetFeatureVector(images[i]);
}
MemoryStream stream = new MemoryStream();
BinaryFormatter fmt = new BinaryFormatter();
fmt.Serialize(stream, bow);
stream.Seek(0, SeekOrigin.Begin);
bow = (BagOfVisualWords <FastRetinaKeypoint, byte[]>)fmt.Deserialize(stream);
double[][] actual = new double[expected.Length][];
for (int i = 0; i < actual.Length; i++)
{
actual[i] = bow.GetFeatureVector(images[i]);
}
Assert.IsTrue(expected.IsEqual(actual));
}
private void trainBoW(object sender, RoutedEventArgs e)
{
//train bow
imageInfo.Text = "Training BoW";
WPFFolderBrowserDialog dd = new WPFFolderBrowserDialog();
dd.Title = "Select a folder";
if (dd.ShowDialog() == true)
{
String[] names = UtilFn.GetFilesFrom(dd.FileName, filtersDir, false);
if (names.Length == 0)
{
MessageBox.Show("NO IMAGES IN SELECTED FOLDER.");
}
ushort c = 0;
List <Bitmap> ims = new List <Bitmap>();
foreach (String fn in names)
{
if (c++ == 2000)
{
break;
}
ims.Add(new Bitmap(fn));
}
bow = new BagOfVisualWords(20); // br features
bow.Learn(ims.ToArray());
}
else
{
MessageBox.Show("Something went wrong.");
}
imageInfo.Text = "Done";
}
public void BagOfVisualWordsConstructorTest3()
{
MoravecCornersDetector moravec = new MoravecCornersDetector();
CornerFeaturesDetector detector = new CornerFeaturesDetector(moravec);
var bow = new BagOfVisualWords <CornerFeaturePoint>(detector, numberOfWords: 10);
var points = bow.Compute(images, 1e-3);
double[] vector = bow.GetFeatureVector(images[0]);
Assert.AreEqual(10, bow.NumberOfWords);
Assert.AreEqual(6, points.Length);
Assert.AreEqual(10, vector.Length);
Assert.AreEqual(2800, points[0].Count);
Assert.AreEqual(4532, points[1].Count);
Assert.AreEqual(2282, points[2].Count);
Assert.AreEqual(1173, points[3].Count);
Assert.AreEqual(4860, points[4].Count);
Assert.AreEqual(5730, points[5].Count);
Assert.AreEqual(596, points[0][0].Descriptor[0]);
Assert.AreEqual(51, points[0][0].Descriptor[1]);
Assert.AreEqual(points[0][0].Descriptor[0], points[0][0].X);
Assert.AreEqual(points[0][0].Descriptor[1], points[0][0].Y);
Assert.AreEqual(461, points[3][7].Descriptor[0]);
Assert.AreEqual(8, points[2][3].Descriptor[1]);
Assert.AreEqual(991, points[2][52].Descriptor[0]);
Assert.AreEqual(82, points[1][11].Descriptor[1]);
Assert.AreEqual(430, points[0][42].Descriptor[0]);
Assert.AreEqual(135, points[4][125].Descriptor[1]);
}
public void GetFeatureVectorTest()
{
Accord.Math.Tools.SetupGenerator(0);
Bitmap image = images[0];
// The Bag-of-Visual-Words model converts arbitrary-size images
// into fixed-length feature vectors. In this example, we will
// be setting the codebook size to 10. This means all generated
// feature vectors will have the same length of 10.
// Create a new Bag-of-Visual-Words (BoW) model
BagOfVisualWords bow = new BagOfVisualWords(10);
// Compute the model using
// a set of training images
bow.Compute(images);
// After this point, we will be able to translate
// images into double[] feature vectors using
double[] feature = bow.GetFeatureVector(image);
Assert.AreEqual(10, feature.Length);
double[][] expected =
{
new double[] { 102, 58, 42, 24, 47, 53, 29, 26, 3, 22 },
new double[] { 90, 135, 71, 101, 26, 28, 105, 61, 62, 48 },
new double[] { 138, 36, 55, 56, 71, 61, 30, 47, 33, 22 }
};
double[][] actual = new double[expected.Length][];
for (int i = 0; i < actual.Length; i++)
actual[i] = bow.GetFeatureVector(images[i]);
// string str = actual.ToString(CSharpJaggedMatrixFormatProvider.InvariantCulture);
for (int i = 0; i < actual.Length; i++)
{
for (int j = 0; j < actual[i].Length; j++)
{
Assert.IsTrue(expected[i].Contains(actual[i][j]));
}
}
}
/// <summary>
/// This methods computes the Bag-of-Visual-Words with the training images.
/// </summary>
///
private void btnBagOfWords_Click(object sender, EventArgs e)
{
int numberOfWords = (int)numWords.Value;
// Create a Binary-Split clustering algorithm
BinarySplit binarySplit = new BinarySplit(numberOfWords);
Stopwatch sw1 = Stopwatch.StartNew();
IBagOfWords<Bitmap> bow;
if (rbSurf.Checked)
{
// Create bag-of-words (BoW) with the given algorithm
var surfBow = new BagOfVisualWords(binarySplit);
// Compute the BoW codebook using training images only
surfBow.Compute(originalTrainImages.Values.ToArray());
bow = surfBow;
}
else
{
// Alternative creation using the FREAK detector
// Create a Binary-Split clustering algorithm
var kmodes = new KModes<byte>(numberOfWords, Distance.BitwiseHamming);
var detector = new FastRetinaKeypointDetector();
// Create bag-of-words (BoW) with the given algorithm
var freakBow = new BagOfVisualWords<FastRetinaKeypoint, byte[]>(detector, kmodes);
// Compute the BoW codebook using training images only
freakBow.Compute(originalTrainImages.Values.ToArray());
bow = freakBow;
}
sw1.Stop();
Stopwatch sw2 = Stopwatch.StartNew();
// Extract features for all images
foreach (ListViewItem item in listView1.Items)
{
// Get item image
Bitmap image = originalImages[item.ImageKey] as Bitmap;
// Process image
double[] featureVector = bow.GetFeatureVector(image);
string featureString = featureVector.ToString(DefaultArrayFormatProvider.InvariantCulture);
if (item.SubItems.Count == 2)
item.SubItems[1].Text = featureString;
else item.SubItems.Add(featureString);
int classLabel = (item.Tag as Tuple<double[], int>).Item2;
item.Tag = Tuple.Create(featureVector, classLabel);
}
sw2.Stop();
lbStatus.Text = "BoW constructed in " + sw1.Elapsed + "s. Features extracted in " + sw2.Elapsed + "s.";
btnSampleRunAnalysis.Enabled = true;
}
public void SerializeTest()
{
Accord.Math.Tools.SetupGenerator(0);
BagOfVisualWords bow = new BagOfVisualWords(10);
bow.Compute(images);
double[][] expected = new double[images.Length][];
for (int i = 0; i < expected.Length; i++)
expected[i] = bow.GetFeatureVector(images[i]);
MemoryStream stream = new MemoryStream();
BinaryFormatter fmt = new BinaryFormatter();
fmt.Serialize(stream, bow);
stream.Seek(0, SeekOrigin.Begin);
bow = (BagOfVisualWords)fmt.Deserialize(stream);
double[][] actual = new double[expected.Length][];
for (int i = 0; i < actual.Length; i++)
actual[i] = bow.GetFeatureVector(images[i]);
Assert.IsTrue(expected.IsEqual(actual));
}
public void GetFeatureVectorTest()
{
Accord.Math.Tools.SetupGenerator(0);
Bitmap image = images[0];
// The Bag-of-Visual-Words model converts arbitrary-size images
// into fixed-length feature vectors. In this example, we will
// be setting the codebook size to 10. This means all generated
// feature vectors will have the same length of 10.
// Create a new Bag-of-Visual-Words (BoW) model
BagOfVisualWords bow = new BagOfVisualWords(10);
// Compute the model using
// a set of training images
bow.Compute(images);
// After this point, we will be able to translate
// images into double[] feature vectors using
double[] feature = bow.GetFeatureVector(image);
Assert.AreEqual(10, feature.Length);
double[][] expected =
{
new double[] { 26, 58, 102, 24, 53, 22, 47, 29, 42, 3 },
new double[] { 61, 135, 90, 101, 28, 48, 26, 105, 71, 62 },
new double[] { 47, 36, 138, 56, 61, 22, 71, 30, 55, 33 },
};
double[][] actual = new double[expected.Length][];
for (int i = 0; i < actual.Length; i++)
actual[i] = bow.GetFeatureVector(images[i]);
Assert.IsTrue(expected.IsEqual(actual));
}
public void BagOfVisualWordsConstructorTest3()
{
MoravecCornersDetector moravec = new MoravecCornersDetector();
CornerFeaturesDetector detector = new CornerFeaturesDetector(moravec);
var bow = new BagOfVisualWords<CornerFeaturePoint>(detector, numberOfWords: 10);
var points = bow.Compute(images, 1e-3);
double[] vector = bow.GetFeatureVector(images[0]);
Assert.AreEqual(10, bow.NumberOfWords);
Assert.AreEqual(6, points.Length);
Assert.AreEqual(10, vector.Length);
Assert.AreEqual(2800, points[0].Count);
Assert.AreEqual(4532, points[1].Count);
Assert.AreEqual(2282, points[2].Count);
Assert.AreEqual(1173, points[3].Count);
Assert.AreEqual(4860, points[4].Count);
Assert.AreEqual(5730, points[5].Count);
Assert.AreEqual(596, points[0][0].Descriptor[0]);
Assert.AreEqual(51, points[0][0].Descriptor[1]);
Assert.AreEqual(points[0][0].Descriptor[0], points[0][0].X);
Assert.AreEqual(points[0][0].Descriptor[1], points[0][0].Y);
Assert.AreEqual(461, points[3][7].Descriptor[0]);
Assert.AreEqual(8, points[2][3].Descriptor[1]);
Assert.AreEqual(991, points[2][52].Descriptor[0]);
Assert.AreEqual(82, points[1][11].Descriptor[1]);
Assert.AreEqual(430, points[0][42].Descriptor[0]);
Assert.AreEqual(135, points[4][125].Descriptor[1]);
}
/// <summary>
/// This methods computes the Bag-of-Visual-Words with the training images.
/// </summary>
///
private void btnBagOfWords_Click(object sender, EventArgs e)
{
int numberOfWords = (int)numWords.Value;
// Create a Binary-Split clustering algorithm
BinarySplit binarySplit = new BinarySplit(numberOfWords);
// Create bag-of-words (BoW) with the given algorithm
BagOfVisualWords bow = new BagOfVisualWords(binarySplit);
if (cbExtended.Checked)
bow.Detector.ComputeDescriptors = SpeededUpRobustFeatureDescriptorType.Extended;
Stopwatch sw1 = Stopwatch.StartNew();
// Compute the BoW codebook using training images only
var points = bow.Compute(originalTrainImages.Values.ToArray());
sw1.Stop();
Stopwatch sw2 = Stopwatch.StartNew();
// Extract features for all images
foreach (ListViewItem item in listView1.Items)
{
// Get item image
Bitmap image = originalImages[item.ImageKey] as Bitmap;
// Process image
double[] featureVector = bow.GetFeatureVector(image);
string featureString = featureVector.ToString(DefaultArrayFormatProvider.InvariantCulture);
if (item.SubItems.Count == 2)
item.SubItems[1].Text = featureString;
else item.SubItems.Add(featureString);
int classLabel = (item.Tag as Tuple<double[], int>).Item2;
item.Tag = Tuple.Create(featureVector, classLabel);
}
sw2.Stop();
lbStatus.Text = "BoW constructed in " + sw1.Elapsed + "s. Features extracted in " + sw2.Elapsed + "s.";
btnSampleRunAnalysis.Enabled = true;
}
public void SerializeTest2()
{
Accord.Math.Tools.SetupGenerator(0);
FastCornersDetector fast = new FastCornersDetector();
FastRetinaKeypointDetector freak = new FastRetinaKeypointDetector(fast);
var kmodes = new KModes<byte[]>(5, Distance.BitwiseHamming);
var bow = new BagOfVisualWords<FastRetinaKeypoint, byte[]>(freak, kmodes);
bow.Compute(images);
double[][] expected = new double[images.Length][];
for (int i = 0; i < expected.Length; i++)
expected[i] = bow.GetFeatureVector(images[i]);
MemoryStream stream = new MemoryStream();
BinaryFormatter fmt = new BinaryFormatter();
fmt.Serialize(stream, bow);
stream.Seek(0, SeekOrigin.Begin);
bow = (BagOfVisualWords<FastRetinaKeypoint, byte[]>)fmt.Deserialize(stream);
double[][] actual = new double[expected.Length][];
for (int i = 0; i < actual.Length; i++)
actual[i] = bow.GetFeatureVector(images[i]);
Assert.IsTrue(expected.IsEqual(actual));
}
public void CalculateFeatureVector(BagOfVisualWords bow)
{
FeatureVector = bow.GetFeatureVector(_image.BitmapImage);
}