/// <summary>
/// Called when videoPlayer receives a new frame.
/// </summary>
/// <param name="sender"></param>
/// <param name="image"></param>
private void videoPlayer_NewFrameReceived(object sender, Accord.Video.NewFrameEventArgs eventArgs)
{
// convert image to dlib format
var img = eventArgs.Frame.ToArray2D <RgbPixel>();
// detect face every 4 frames
if (frameIndex % 4 == 0)
{
var faces = faceDetector.Detect(img);
if (faces.Length > 0)
{
currentFace = faces.First();
}
}
// abort if we don't have a face at this point
if (currentFace == default(DlibDotNet.Rectangle))
{
return;
}
// detect facial landmarks
var shape = shapePredictor.Detect(img, currentFace);
// detect head pose
if (shape.Parts == 68)
{
DetectHeadPose(eventArgs.Frame, shape);
}
// update frame counter
frameIndex++;
}
/// <summary>
/// Called when videoPlayer receives a new frame.
/// </summary>
/// <param name="sender"></param>
/// <param name="image"></param>
private void videoPlayer_NewFrame(object sender, ref System.Drawing.Bitmap image)
{
// conver frame to grayscale
var grayscale = new GrayscaleBT709();
var grayImage = grayscale.Apply(image);
// convert image to dlib format
var img = grayImage.ToArray2D <RgbPixel>();
// detect face every 4 frames
if (frameIndex % 4 == 0)
{
var faces = faceDetector.Detect(img);
if (faces.Length > 0)
{
currentFace = faces.First();
}
}
// abort if we don't have a face at this point
if (currentFace == default(DlibDotNet.Rectangle))
{
return;
}
// detect facial landmarks
var shape = shapePredictor.Detect(img, currentFace);
// detect eye state
DetectEyeState(image, shape);
// update frame counter
frameIndex++;
}
private void CheckFace(double picth, Mat frame, Rectangle face, double yaw, double pitch)
{
if (this.countdown == 0)
{
this.picture.Image = frame.ToBitmap();
this.SuccessMsg.Visible = true;
SetStart();
SetZero();
}
else if (this.step == this.checker.Length)
{
this.stopwatch.Start();
CountDown(face);
if (!IsForntFace(yaw, pitch))
{
SetZero();
this.ErrorMsg.Visible = true;
}
}
else if (this.step == 0 && IsFaceInFrame(face) || Math.Abs(this.checker[this.step] - picth) <= 5)
{
checkedListBox.SetItemChecked(this.step, true);
this.step++;
}
}
private void CountDown(Rectangle face)
{
var time = this.stopwatch.Elapsed.TotalSeconds;
this.countdown = this.timeset - (int)this.stopwatch.Elapsed.TotalSeconds;
checkedListBox.Items.RemoveAt(3);
checkedListBox.Items.Insert(3, $"4. มองตรง ({this.countdown} วินาที)");
checkedListBox.SetItemChecked(3, true);
}
private static List <(OutputLabels <Matrix <float> >, Rectangle[])> GetData(List <Bitmap> bitmaps, bool isAFace = false)
{
var datas = new List <(OutputLabels <Matrix <float> >, Rectangle[])>();
try
{
foreach (var bitmap in bitmaps)
{
var faces = new List <Matrix <RgbPixel> >();
var dets = new Rectangle[0];
//在图像中寻找人脸我们需要一个人脸检测器:
using (var detector = Dlib.GetFrontalFaceDetector())
{
using (var img = bitmap.ToMatrix <RgbPixel>())
{
// 人脸 面积从大到小排序
dets = detector.Operator(img).OrderByDescending(x => x.Area).ToArray();
// 是否只检测面积最大的人脸
if (isAFace)
{
var shape = _SP.Detect(img, dets[0]);
var faceChipDetail = Dlib.GetFaceChipDetails(shape, 150, 0.25);
var faceChip = Dlib.ExtractImageChip <RgbPixel>(img, faceChipDetail);
faces.Add(faceChip);
}
else
{
foreach (var face in dets)
{
var shape = _SP.Detect(img, face);
var faceChipDetail = Dlib.GetFaceChipDetails(shape, 150, 0.25);
var faceChip = Dlib.ExtractImageChip <RgbPixel>(img, faceChipDetail);
faces.Add(faceChip);
}
}
if (!faces.Any())
{
datas.Add((null, null));
}
else
{
//此调用要求DNN将每个人脸图像转换为128D矢量。
//在这个128D的矢量空间中，来自同一个人的图像会彼此接近
//但是来自不同人的载体将会非常不同。所以我们可以用这些向量
//辨别一对图片是来自同一个人还是不同的人。
datas.Add((_NET.Operator(faces), dets));
}
}
}
}
}
catch (Exception ex)
{
LogHelperNLog.Error(ex);
}
return(datas);
}
/// <summary>
/// Adjust the rectangle to make sure it is inside the range of the image
/// </summary>
/// <param name="rect">rectangle to be adjusted</param>
/// <param name="img">image for reference</param>
/// <returns></returns>
public static DlibDotNet.Rectangle RectangleAdjust(DlibDotNet.Rectangle rect, Array2D <RgbPixel> img)
{
DlibDotNet.Rectangle fitRect = new DlibDotNet.Rectangle();
fitRect.Right = rect.Right < img.Rect.Right ? rect.Right : img.Rect.Right;
fitRect.Left = rect.Left > img.Rect.Left ? rect.Left : img.Rect.Left;
fitRect.Top = rect.Top > img.Rect.Top ? rect.Top : img.Rect.Top;
fitRect.Bottom = rect.Bottom < img.Rect.Bottom ? rect.Bottom : img.Rect.Bottom;
return(fitRect);
}
/// <summary>
/// 具体计算
/// </summary>
/// <param name="bitmap"></param>
/// <returns></returns>
public Rectangle[] Face(Bitmap bitmap)
{
var dets = new Rectangle[0];
using (var detector = Dlib.GetFrontalFaceDetector())
//using (var img = Dlib.LoadImage<RgbPixel>("png.png"))
using (var img = bitmap.ToArray2D <RgbPixel>())
{
dets = detector.Operator(img);
}
return(dets);
}
private DlibDotNet.Rectangle ConvertToDlib(Rect rect)
{
DlibDotNet.Rectangle dlibRect = new DlibDotNet.Rectangle()
{
Left = rect.X,
Top = rect.Y,
Right = rect.X + rect.Width,
Bottom = rect.Y + rect.Height
};
return(dlibRect);
}
private bool IsFaceInFrame(Rectangle face)
{
var percent = (double)(face.Area) / (this.size.Width * this.size.Height);
var isFaceInFrame = this.rect.Left < face.Center.X && face.Center.X < this.rect.Right &&
this.rect.Top < face.Center.Y && face.Center.Y < this.rect.Bottom &&
-0.5 < percent - 1 && percent - 1 < 1;
this.ErrorMsg.Visible = !isFaceInFrame;
return(isFaceInFrame);
}
/// <summary>
/// 使用路径获取位置数据
/// </summary>
/// <param name="url"></param>
/// <returns></returns>
public static Rectangle[] GetResult(string url)
{
var dets = new Rectangle[0];
url = Path.Combine(AppDomain.CurrentDomain.BaseDirectory, url);
using (var detector = Dlib.GetFrontalFaceDetector())
//using (var img = Dlib.LoadImage<RgbPixel>("png.png"))
using (var img = Dlib.LoadImage <RgbPixel>(url))
{
dets = detector.Operator(img);
}
return(dets);
}
/// <summary>
/// Called when videoPlayer receives a new frame.
/// </summary>
/// <param name="sender"></param>
/// <param name="image"></param>
private void videoPlayer_NewFrame(object sender, ref Bitmap image)
{
// convert image to dlib format
var img = image.ToArray2D <RgbPixel>();
// find the face
// note that we only detect faces every 4 frames
if (faceRect == default(DlibDotNet.Rectangle) || frameIndex++ % 4 == 0)
{
var faces = faceDetector.Detect(img);
faceRect = faces.FirstOrDefault();
}
// abort if we found no face
if (faceRect == default(DlibDotNet.Rectangle))
{
return;
}
// find face landmark points
var shape = shapePredictor.Detect(img, faceRect);
var landmarkPoints = BeardHelper.GetLandmarkPoints(shape);
// find beard landmark points
var beardPoints = BeardHelper.GetBeardPoints();
// calculate Delaunay triangles
var triangles = Utility.GetDelaunayTriangles(landmarkPoints);
// get transformations to warp the beard onto the face
var warps = Utility.GetWarps(beardPoints, landmarkPoints, triangles);
// split the beard image into an alpha mask and an RGB part
var beard = BitmapConverter.ToMat(beardImage);
BeardHelper.SplitChannels(beard, out var beardMask, out var beardRgb);
// warp the beard RGB image
var warpedBeard = Utility.ApplyWarps(BitmapConverter.ToBitmap(beardRgb), image.Width, image.Height, warps);
// warp the beard alpha mask
var warpedBeardMask = Utility.ApplyWarps(BitmapConverter.ToBitmap(beardMask), image.Width, image.Height, warps);
// blend the beard onto the camera frame by using the mask
var frame = BitmapConverter.ToMat(image);
var result = BeardHelper.Blend(warpedBeard, warpedBeardMask, frame);
// return result
image = BitmapConverter.ToBitmap(result);
}
protected override void Demo(FaceRecognition faceRecognition, string modelFile, string imageFile, Image image, Location location)
{
var networkId = SetupNetwork();
using (var net = LossMulticlassLog.Deserialize(modelFile, networkId))
using (var bitmap = (Bitmap)System.Drawing.Image.FromFile(imageFile))
using (var org = new Bitmap(bitmap.Width, bitmap.Height))
using (var g = Graphics.FromImage(org))
{
g.DrawImage(bitmap, new System.Drawing.Rectangle(0, 0, org.Width, org.Height), new System.Drawing.Rectangle(0, 0, bitmap.Width, bitmap.Height), GraphicsUnit.Pixel);
var rect = new Rectangle(location.Left, location.Top, location.Right, location.Bottom);
var dPoint = new[]
{
new DPoint(rect.Left, rect.Top),
new DPoint(rect.Right, rect.Top),
new DPoint(rect.Left, rect.Bottom),
new DPoint(rect.Right, rect.Bottom),
};
using (var tmp = Dlib.LoadImageAsMatrix <byte>(imageFile))
using (var face = Dlib.ExtractImage4Points(tmp, dPoint, this.Size, this.Size))
{
this.SetEvalMode(networkId, net);
var results = net.Probability(face, 1).ToArray();
var labels = net.GetLabels();
var dictionary = new Dictionary <string, float>();
for (var index = 0; index < labels.Length; index++)
{
dictionary.Add(labels[index], results[0][index]);
}
var maxResult = dictionary.Aggregate((max, working) => (max.Value > working.Value) ? max : working);
var emotion = maxResult.Key;
var probability = maxResult.Value;
using (var p = new Pen(Color.Red, bitmap.Width / 200f))
using (var b = new SolidBrush(Color.Blue))
using (var font = new Font("Calibri", 16))
{
g.DrawRectangle(p, rect.Left, rect.Top, rect.Width, rect.Height);
g.DrawString($"{emotion}\n({probability})", font, b, new PointF(rect.Left + 10, rect.Top + 10));
}
org.Save("demo.jpg");
}
}
}
private static byte[] ExtractFace(Rect face, Bitmap source)
{
using (var target = new Bitmap((int)face.Width, (int)face.Height))
{
using (Graphics g = Graphics.FromImage(target))
{
g.DrawImage(source, new Rectangle(0, 0, target.Width, target.Height), new Rectangle(face.Left, face.Top, (int)face.Width, (int)face.Height), GraphicsUnit.Pixel);
}
using (var memoryStream = new MemoryStream())
{
target.Save(memoryStream, ImageFormat.Png);
return(memoryStream.ToArray());
}
}
}
private void OnCameraFrame(object sender, EventArgs e)
{
img = capture.RetrieveMat();
Cv2.Flip(img, img, FlipMode.Y);
var array = new byte[img.Cols * img.Rows * img.ElemSize()];
Marshal.Copy(img.Data, array, 0, array.Length);
var image = Dlib.LoadImageData <RgbPixel>(array, (uint)img.Rows, (uint)img.Cols, (uint)(img.Cols * img.ElemSize()));
faces = detector.Operator(image);
shapes.Clear();
foreach (var rect in faces)
{
DlibDotNet.Rectangle face = rect;
shapes.Add(predictor.Detect(image, face));
}
Invalidate();
}
/// <summary>
/// Detect all 68 landmarks on the face on camera
/// </summary>
/// <param name="image">The current camera frame to analyze</param>
/// <param name="frameIndex">The index number of the current camera frame</param>
/// <returns>A FullObjectDetection object containing all 68 facial landmark points</returns>
private FullObjectDetection DetectLandmarks(Bitmap image, int frameIndex)
{
// convert image to dlib format
var dlibImage = image.ToArray2D <RgbPixel>();
// detect faces every 5 frames
if (frameIndex % 5 == 0)
{
var faces = faceDetector.Detect(dlibImage);
if (faces.Length > 0)
{
// grab the first face
currentFace = faces.First();
}
}
// detect all 68 facial landmarks on the face
if (currentFace != default(DlibDotNet.Rectangle))
{
return(shapePredictor.Detect(dlibImage, currentFace));
}
return(null);
}
private void Timer1_Tick(object sender, EventArgs e)
{
capture.Read(frame);
this.point = new Point((frame.Width - size.Width) / 2, (frame.Height - size.Height) / 2);
this.rect = new Rect(point, size);
Cv2.Flip(frame, frame, FlipMode.Y);
if (!frame.Empty() && start)
{
var img = ConvertToArray2D(frame);
var faces = fd.Operator(img);
if (faces.Any(face => IsFaceInFrame(face)))
{
foreach (var face in faces)
{
if (IsFaceInFrame(face))
{
//Dlib.DrawRectangle(img, face, color: new RgbPixel(0, 255, 255), thickness: 4);
var shape = sp.Detect(img, face);
var landmarks = new MatOfPoint2d(1, 6,
(from i in new int[] { 30, 8, 36, 45, 48, 54 }
let pt = shape.GetPart((uint)i)
select new OpenCvSharp.Point2d(pt.X, pt.Y)).ToArray());
var cameraMatrix = Utility.GetCameraMatrix((int)img.Rect.Width, (int)img.Rect.Height);
Mat rotation = new MatOfDouble();
Mat translation = new MatOfDouble();
Cv2.SolvePnP(model, landmarks, cameraMatrix, coeffs, rotation, translation);
var euler = Utility.GetEulerMatrix(rotation);
var yaw = 180 * euler.At <double>(0, 2) / Math.PI;
var pitch = 180 * euler.At <double>(0, 1) / Math.PI;
pitch = Math.Sign(pitch) * 180 - pitch;
Cv2.ProjectPoints(poseModel, rotation, translation, cameraMatrix, coeffs, poseProjection);
//var landmark = landmarks.At<Point2d>(0);
//var p = poseProjection.At<Point2d>(0);
//Dlib.DrawLine(
// img,
// new DlibDotNet.Point((int)landmark.X, (int)landmark.Y),
// new DlibDotNet.Point((int)p.X, (int)p.Y),
// color: new RgbPixel(0, 255, 255));
//foreach (var i in new int[] { 30, 8, 36, 45, 48, 54 })
//{
// var point = shape.GetPart((uint)i);
// var rect = new Rectangle(point);
// Dlib.DrawRectangle(img, rect, color: new RgbPixel(255, 255, 0), thickness: 4);
//}
for (var i = 0; i < shape.Parts; i++)
{
var point = shape.GetPart((uint)i);
var rect = new Rectangle(point);
Dlib.DrawRectangle(img, rect, color: new RgbPixel(0, 255, 255), thickness: 4);
}
CheckFace(pitch, frame, face, yaw, pitch);
frame = img.ToBitmap().ToMat();
}
}
}
else if (this.step > 0)
{
SetZero();
this.ErrorMsg.Visible = true;
}
}
Cv2.Rectangle(frame, rect, Scalar.Yellow, thickness: 2);
camera.Image = frame.ToBitmap();
}
private static Location TrimBound(Rectangle location, int width, int height)
{
return(new Location(Math.Max(location.Left, 0), Math.Max(location.Top, 0), Math.Min(location.Right, width), Math.Min(location.Bottom, height)));
}
public int Start(string[] args)
{
var app = new CommandLineApplication(false);
app.Name = this._Name;
app.Description = this._Description;
app.HelpOption("-h|--help");
app.Command("clean", command =>
{
var outputOption = command.Option("-o|--output", "The output directory path.", CommandOptionType.SingleValue);
command.OnExecute(() =>
{
if (!outputOption.HasValue() || !Directory.Exists(outputOption.Value()))
{
Logger.Error($"'{outputOption.Value()} is missing or output option is not specified");
return(-1);
}
Logger.Info($" Output: {outputOption.Value()}");
Logger.Info("");
Clean(outputOption.Value());
return(0);
});
});
app.Command("train", command =>
{
const uint epochDefault = 300;
const double learningRateDefault = 0.001d;
const double minLearningRateDefault = 0.00001d;
const uint minBatchSizeDefault = 256;
const uint validationDefault = 30;
var datasetOption = command.Option("-d|--dataset", "The directory of dataset", CommandOptionType.SingleValue);
var epochOption = command.Option("-e|--epoch", $"The epoch. Default is {epochDefault}", CommandOptionType.SingleValue);
var learningRateOption = command.Option("-l|--lr", $"The learning rate. Default is {learningRateDefault}", CommandOptionType.SingleValue);
var minLearningRateOption = command.Option("-m|--min-lr", $"The minimum learning rate. Default is {minLearningRateDefault}", CommandOptionType.SingleValue);
var minBatchSizeOption = command.Option("-b|--min-batchsize", $"The minimum batch size. Default is {minBatchSizeDefault}", CommandOptionType.SingleValue);
var validationOption = command.Option("-v|--validation-interval", $"The interval of validation. Default is {validationDefault}", CommandOptionType.SingleValue);
var useMeanOption = command.Option("-u|--use-mean", "Use mean image", CommandOptionType.NoValue);
var outputOption = command.Option("-o|--output", "The output directory path.", CommandOptionType.SingleValue);
command.OnExecute(() =>
{
var dataset = datasetOption.Value();
if (!datasetOption.HasValue() || !Directory.Exists(dataset))
{
Logger.Error("dataset does not exist");
return(-1);
}
var epoch = epochDefault;
if (epochOption.HasValue() && !uint.TryParse(epochOption.Value(), out epoch))
{
Logger.Error("epoch is invalid value");
return(-1);
}
var learningRate = learningRateDefault;
if (learningRateOption.HasValue() && !double.TryParse(learningRateOption.Value(), NumberStyles.Float, Thread.CurrentThread.CurrentCulture.NumberFormat, out learningRate))
{
Logger.Error("learning rate is invalid value");
return(-1);
}
var minLearningRate = minLearningRateDefault;
if (minLearningRateOption.HasValue() && !double.TryParse(minLearningRateOption.Value(), NumberStyles.Float, Thread.CurrentThread.CurrentCulture.NumberFormat, out minLearningRate))
{
Logger.Error("minimum learning rate is invalid value");
return(-1);
}
var minBatchSize = minBatchSizeDefault;
if (minBatchSizeOption.HasValue() && !uint.TryParse(minBatchSizeOption.Value(), out minBatchSize))
{
Logger.Error("minimum batch size is invalid value");
return(-1);
}
var validation = validationDefault;
if (validationOption.HasValue() && !uint.TryParse(validationOption.Value(), out validation) || validation == 0)
{
Logger.Error("validation interval is invalid value");
return(-1);
}
var output = "result";
if (outputOption.HasValue())
{
output = outputOption.Value();
}
Directory.CreateDirectory(output);
var useMean = useMeanOption.HasValue();
Logger.Info($" Dataset: {dataset}");
Logger.Info($" Epoch: {epoch}");
Logger.Info($" Learning Rate: {learningRate}");
Logger.Info($" Min Learning Rate: {minLearningRate}");
Logger.Info($" Min Batch Size: {minBatchSize}");
Logger.Info($"Validation Interval: {validation}");
Logger.Info($" Use Mean: {useMean}");
Logger.Info($" Output: {output}");
Logger.Info("");
var name = this.GetBaseName(epoch, learningRate, minLearningRate, minBatchSize);
var baseName = Path.Combine(output, name);
var parameter = new Parameter
{
BaseName = baseName,
Dataset = dataset,
Output = output,
Epoch = epoch,
LearningRate = learningRate,
MinLearningRate = minLearningRate,
MiniBatchSize = minBatchSize,
Validation = validation
};
Train(parameter);
return(0);
});
});
app.Command("test", command =>
{
var datasetOption = command.Option("-d|--dataset", "The directory of dataset", CommandOptionType.SingleValue);
var modelOption = command.Option("-m|--model", "The model file path", CommandOptionType.SingleValue);
command.OnExecute(() =>
{
var dataset = datasetOption.Value();
if (!datasetOption.HasValue() || !Directory.Exists(dataset))
{
Logger.Error("dataset does not exist");
return(-1);
}
var model = modelOption.Value();
if (!modelOption.HasValue() || !File.Exists(model))
{
Logger.Error("model does not exist");
return(-1);
}
Logger.Info($"Dataset: {dataset}");
Logger.Info($" Model: {model}");
Logger.Info("");
var parameter = new Parameter
{
Dataset = dataset,
Model = model
};
Test(parameter);
return(0);
});
});
app.Command("eval", command =>
{
var imageOption = command.Option("-i|--image", "The image file.", CommandOptionType.SingleValue);
var modelOption = command.Option("-m|--model", "The model file path", CommandOptionType.SingleValue);
command.OnExecute(() =>
{
var image = imageOption.Value();
if (!imageOption.HasValue() || !File.Exists(image))
{
Logger.Error("image does not exist");
return(-1);
}
var model = modelOption.Value();
if (!modelOption.HasValue() || !File.Exists(model))
{
Logger.Error("model file does not exist");
return(-1);
}
Logger.Info($"Image File: {image}");
Logger.Info($" Model: {model}");
Logger.Info("");
var networkId = SetupNetwork();
using (var net = LossMulticlassLog.Deserialize(model, networkId))
using (var fr = FaceRecognition.Create("Models"))
using (var img = FaceRecognition.LoadImageFile(image))
{
var location = fr.FaceLocations(img).FirstOrDefault();
if (location == null)
{
Logger.Info("Missing face");
return(-1);
}
var rect = new Rectangle(location.Left, location.Top, location.Right, location.Bottom);
var dPoint = new[]
{
new DPoint(rect.Left, rect.Top),
new DPoint(rect.Right, rect.Top),
new DPoint(rect.Left, rect.Bottom),
new DPoint(rect.Right, rect.Bottom),
};
using (var tmp = Dlib.LoadImageAsMatrix <byte>(image))
{
using (var face = Dlib.ExtractImage4Points(tmp, dPoint, this.Size, this.Size))
{
this.SetEvalMode(networkId, net);
using (var predictedLabels = net.Operator(face))
Logger.Info($"{this.Cast(predictedLabels[0])}");
}
}
}
return(0);
});
});
app.Command("demo", command =>
{
command.HelpOption("-?|-h|--help");
var imageOption = command.Option("-i|--image", "test image file", CommandOptionType.SingleValue);
var modelOption = command.Option("-m|--model", "model file", CommandOptionType.SingleValue);
var directoryOption = command.Option("-d|--directory", "model files directory path", CommandOptionType.SingleValue);
command.OnExecute(() =>
{
if (!imageOption.HasValue())
{
Console.WriteLine("image option is missing");
app.ShowHelp();
return(-1);
}
if (!directoryOption.HasValue())
{
Console.WriteLine("directory option is missing");
app.ShowHelp();
return(-1);
}
if (!modelOption.HasValue())
{
Console.WriteLine("model option is missing");
app.ShowHelp();
return(-1);
}
var modelFile = modelOption.Value();
if (!File.Exists(modelFile))
{
Console.WriteLine($"'{modelFile}' is not found");
app.ShowHelp();
return(-1);
}
var imageFile = imageOption.Value();
if (!File.Exists(imageFile))
{
Console.WriteLine($"'{imageFile}' is not found");
app.ShowHelp();
return(-1);
}
var directory = directoryOption.Value();
if (!Directory.Exists(directory))
{
Console.WriteLine($"'{directory}' is not found");
app.ShowHelp();
return(-1);
}
using (var fr = FaceRecognition.Create(directory))
using (var image = FaceRecognition.LoadImageFile(imageFile))
{
var loc = fr.FaceLocations(image).FirstOrDefault();
if (loc == null)
{
Console.WriteLine("No face is detected");
return(0);
}
this.Demo(fr, modelFile, imageFile, image, loc);
}
return(0);
});
});
return(app.Execute(args));
}
private static void Main(string[] args)
{
var app = new CommandLineApplication(false);
app.Name = nameof(HelenTraining);
app.Description = "The program for training helen dataset";
app.HelpOption("-h|--help");
app.Command("generate", command =>
{
command.HelpOption("-?|-h|--help");
var paddingOption = command.Option("-p|--padding", "padding of detected face", CommandOptionType.SingleValue);
var modelsOption = command.Option("-m|--model", "model files directory path", CommandOptionType.SingleValue);
command.OnExecute(() =>
{
if (!modelsOption.HasValue())
{
Console.WriteLine("model option is missing");
app.ShowHelp();
return(-1);
}
if (!paddingOption.HasValue())
{
Console.WriteLine("padding option is missing");
app.ShowHelp();
return(-1);
}
var directory = modelsOption.Value();
if (!Directory.Exists(directory))
{
Console.WriteLine($"'{directory}' is not found");
app.ShowHelp();
return(-1);
}
if (!int.TryParse(paddingOption.Value(), out var padding))
{
Console.WriteLine($"padding '{paddingOption.Value()}' is not integer");
app.ShowHelp();
return(-1);
}
Console.WriteLine($"Model: {directory}");
Console.WriteLine($"Padding: {padding}");
_FaceRecognition = FaceRecognition.Create(directory);
const string extractPath = "helen";
var zips = new[]
{
new{ Zip = "annotation.zip", IsImage = false, Directory = "annotation" },
new{ Zip = "helen_1.zip", IsImage = true, Directory = "helen_1" },
new{ Zip = "helen_2.zip", IsImage = true, Directory = "helen_2" },
new{ Zip = "helen_3.zip", IsImage = true, Directory = "helen_3" },
new{ Zip = "helen_4.zip", IsImage = true, Directory = "helen_4" },
new{ Zip = "helen_5.zip", IsImage = true, Directory = "helen_5" }
};
Directory.CreateDirectory(extractPath);
foreach (var zip in zips)
{
if (!Directory.Exists(Path.Combine(extractPath, zip.Directory)))
{
ZipFile.ExtractToDirectory(zip.Zip, extractPath);
}
}
var annotation = zips.FirstOrDefault(arg => !arg.IsImage);
var imageZips = zips.Where(arg => arg.IsImage).ToArray();
if (annotation == null)
{
return(-1);
}
var images = new List <Image>();
foreach (var file in Directory.EnumerateFiles(Path.Combine(extractPath, annotation.Directory)))
{
Console.WriteLine($"Process: '{file}'");
var txt = File.ReadAllLines(file);
var filename = txt[0];
var jpg = $"{filename}.jpg";
foreach (var imageZip in imageZips)
{
var found = false;
var path = Path.Combine(Path.Combine(extractPath, imageZip.Directory, jpg));
if (File.Exists(path))
{
found = true;
using (var fi = FaceRecognition.LoadImageFile(path))
{
var locations = _FaceRecognition.FaceLocations(fi, 1, Model.Hog).ToArray();
if (locations.Length != 1)
{
Console.WriteLine($"\t'{path}' has {locations.Length} faces.");
}
else
{
var location = locations.First();
var parts = new List <Part>();
for (var i = 1; i < txt.Length; i++)
{
var tmp = txt[i].Split(',').Select(s => s.Trim()).Select(float.Parse).Select(s => (int)s).ToArray();
parts.Add(new Part {
X = tmp[0], Y = tmp[1], Name = $"{i - 1}"
});
}
var image = new Image
{
File = Path.Combine(imageZip.Directory, jpg),
Box = new Box
{
Left = location.Left - padding,
Top = location.Top - padding,
Width = location.Right - location.Left + 1 + padding * 2,
Height = location.Bottom - location.Top + 1 + padding * 2,
Part = parts.ToArray()
}
};
using (var bitmap = System.Drawing.Image.FromFile(path))
{
var b = image.Box;
using (var g = Graphics.FromImage(bitmap))
{
using (var p = new Pen(Color.Red, bitmap.Width / 400f))
g.DrawRectangle(p, b.Left, b.Top, b.Width, b.Height);
foreach (var part in b.Part)
{
g.FillEllipse(Brushes.GreenYellow, part.X, part.Y, 5, 5);
}
}
var result = Path.Combine(extractPath, "Result");
Directory.CreateDirectory(result);
bitmap.Save(Path.Combine(result, jpg), ImageFormat.Jpeg);
}
images.Add(image);
}
}
}
if (found)
{
break;
}
}
}
var dataset = new Dataset
{
Name = "helen dataset",
Comment = "Created by Takuya Takeuchi.",
Images = images.ToArray()
};
var settings = new XmlWriterSettings();
using (var sw = new StreamWriter(Path.Combine(extractPath, "helen-dataset.xml"), false, new System.Text.UTF8Encoding(false)))
using (var writer = XmlWriter.Create(sw, settings))
{
writer.WriteProcessingInstruction("xml-stylesheet", @"type=""text/xsl"" href=""image_metadata_stylesheet.xsl""");
var serializer = new XmlSerializer(typeof(Dataset));
serializer.Serialize(writer, dataset);
}
return(0);
});
});
app.Command("train", command =>
{
command.HelpOption("-?|-h|--help");
var threadOption = command.Option("-t|--threads", "number of threads", CommandOptionType.SingleValue);
var xmlOption = command.Option("-x|--xml", "generated xml file from helen dataset", CommandOptionType.SingleValue);
command.OnExecute(() =>
{
if (!xmlOption.HasValue())
{
Console.WriteLine("xml option is missing");
app.ShowHelp();
return(-1);
}
if (!threadOption.HasValue())
{
Console.WriteLine("thread option is missing");
app.ShowHelp();
return(-1);
}
var xmlFile = xmlOption.Value();
if (!File.Exists(xmlFile))
{
Console.WriteLine($"'{xmlFile}' is not found");
app.ShowHelp();
return(-1);
}
if (!uint.TryParse(threadOption.Value(), out var thread))
{
Console.WriteLine($"thread '{threadOption.Value()}' is not integer");
app.ShowHelp();
return(-1);
}
Dlib.LoadImageDataset(xmlFile, out Array <Array2D <byte> > imagesTrain, out var facesTrain);
using (var trainer = new ShapePredictorTrainer())
{
trainer.NumThreads = thread;
trainer.BeVerbose();
Console.WriteLine("Start training");
using (var predictor = trainer.Train(imagesTrain, facesTrain))
{
Console.WriteLine("Finish training");
var directory = Path.GetDirectoryName(xmlFile);
var output = Path.Combine(directory, $"{Path.GetFileNameWithoutExtension(xmlFile)}.dat");
ShapePredictor.Serialize(predictor, output);
}
}
return(0);
});
});
app.Command("demo", command =>
{
command.HelpOption("-?|-h|--help");
var imageOption = command.Option("-i|--image", "test image file", CommandOptionType.SingleValue);
var modelOption = command.Option("-m|--model", "model file", CommandOptionType.SingleValue);
var directoryOption = command.Option("-d|--directory", "model files directory path", CommandOptionType.SingleValue);
command.OnExecute(() =>
{
if (!imageOption.HasValue())
{
Console.WriteLine("image option is missing");
app.ShowHelp();
return(-1);
}
if (!directoryOption.HasValue())
{
Console.WriteLine("directory option is missing");
app.ShowHelp();
return(-1);
}
if (!modelOption.HasValue())
{
Console.WriteLine("model option is missing");
app.ShowHelp();
return(-1);
}
var modelFile = modelOption.Value();
if (!File.Exists(modelFile))
{
Console.WriteLine($"'{modelFile}' is not found");
app.ShowHelp();
return(-1);
}
var imageFile = imageOption.Value();
if (!File.Exists(imageFile))
{
Console.WriteLine($"'{imageFile}' is not found");
app.ShowHelp();
return(-1);
}
var directory = directoryOption.Value();
if (!Directory.Exists(directory))
{
Console.WriteLine($"'{directory}' is not found");
app.ShowHelp();
return(-1);
}
_FaceRecognition = FaceRecognition.Create(directory);
using (var predictor = ShapePredictor.Deserialize(modelFile))
using (var image = FaceRecognition.LoadImageFile(imageFile))
using (var mat = Dlib.LoadImageAsMatrix <RgbPixel>(imageFile))
using (var bitmap = (Bitmap)System.Drawing.Image.FromFile(imageFile))
using (var white = new Bitmap(bitmap.Width, bitmap.Height))
using (var g = Graphics.FromImage(bitmap))
using (var gw = Graphics.FromImage(white))
{
var loc = _FaceRecognition.FaceLocations(image).FirstOrDefault();
if (loc == null)
{
Console.WriteLine("No face is detected");
return(0);
}
var b = new DlibDotNet.Rectangle(loc.Left, loc.Top, loc.Right, loc.Bottom);
var detection = predictor.Detect(mat, b);
using (var p = new Pen(Color.Red, bitmap.Width / 200f))
{
g.DrawRectangle(p, loc.Left, b.Top, b.Width, b.Height);
gw.Clear(Color.White);
gw.DrawRectangle(p, loc.Left, b.Top, b.Width, b.Height);
}
for (int i = 0, parts = (int)detection.Parts; i < parts; i++)
{
var part = detection.GetPart((uint)i);
g.FillEllipse(Brushes.GreenYellow, part.X, part.Y, 15, 15);
gw.DrawString($"{i}", SystemFonts.DefaultFont, Brushes.Black, part.X, part.Y);
}
bitmap.Save("demo.jpg", ImageFormat.Jpeg);
white.Save("white.jpg", ImageFormat.Jpeg);
}
return(0);
});
});
app.Execute(args);
}
public static void CreateFeatureVectors()
{
int faceCount = 0;
float leftEyebrow, rightEyebrow, leftLip, rightLip, lipHeight, lipWidth;
string output;
if (currentDataType == Datatype.Testing)
{
output = testingOutput;
}
else
{
output = trainingOutput;
}
string[] dirs = Directory.GetFiles(currentFilePath, "*.*", SearchOption.AllDirectories);
// Set up Dlib Face Detector
using (var fd = Dlib.GetFrontalFaceDetector())
// ... and Dlib Shape Detector
using (var sp = ShapePredictor.Deserialize("shape_predictor_68_face_landmarks.dat"))
{
string header = "leftEyebrow,rightEyebrow,leftLip,rightLip,lipWidth,lipHeight,label\n";
// Create the CSV file and fill in the first line with the header
System.IO.File.WriteAllText(output, header);
foreach (string dir in dirs)
{
// call function that sets the label based on what the filename contains
string label = DetermineLabel(dir);
// load input image
if (!(dir.EndsWith("png") || dir.EndsWith("jpg")))
{
continue;
}
var img = Dlib.LoadImage <RgbPixel>(dir);
// find all faces in the image
var faces = fd.Operator(img);
// for each face draw over the facial landmarks
foreach (var face in faces)
{
// Write to the console displaying the progress and current emotion
Form1.SetProgress(faceCount, dirs.Length - 1);
// find the landmark points for this face
var shape = sp.Detect(img, face);
for (var i = 0; i < shape.Parts; i++)
{
RgbPixel colour = new RgbPixel(255, 255, 255);
var point = shape.GetPart((uint)i);
var rect = new DlibDotNet.Rectangle(point);
Dlib.DrawRectangle(img, rect, color: colour, thickness: 2);
}
SetFormImage(img);
leftEyebrow = CalculateLeftEyebrow(shape);
rightEyebrow = CalculateRightEyebrow(shape);
leftLip = CalculateLeftLip(shape);
rightLip = CalculateRightLip(shape);
lipWidth = CalculateLipWidth(shape);
lipHeight = CalculateLipHeight(shape);
using (System.IO.StreamWriter file = new System.IO.StreamWriter(output, true))
{
file.WriteLine(leftEyebrow + "," + rightEyebrow + "," + leftLip + "," + rightLip + "," + lipWidth + "," + lipHeight + "," + label);
}
// Increment count used for console output
faceCount++;
}
}
if (currentDataType == Datatype.Testing)
{
var testDataView = mlContext.Data.LoadFromTextFile <FeatureInputData>(output, hasHeader: true, separatorChar: ',');
GenerateMetrics(testDataView);
}
Form1.HideImage();
}
}
