/// <summary>
/// Find object in darknet yolo
/// </summary>
/// <param name="image">input image</param>
/// <param name="threshold">minimum threshold</param>
/// <returns>Image with labels, labels with probability, runtime, json</returns>
public Task <Tuple <Mat, List <string>, long, string> > FindObjects(Mat image, double threshold)
{
return(Task.Run(() =>
{
// read size from cfg file
var size = int.Parse(
new string(File.ReadAllLines(_cfg).First(x => x.StartsWith("width"))
.Where(char.IsDigit).ToArray())
);
// setting blob, parameter are important
Mat blob = CvDnn.BlobFromImage(image, 1 / 255.0, new Size(size, size),
new Scalar(), true, false);
_net.SetInput(blob, "data");
var sw = new Stopwatch();
sw.Start();
// forward model
Mat prob = _net.Forward();
sw.Stop();
var result = new List <string>();
long time = sw.ElapsedMilliseconds;
const int prefix = 5;
var objectNumbers = new Dictionary <int, int>();
var objects = new List <object>();
for (int i = 0; i < prob.Rows; i++)
{
var confidence = prob.At <float>(i, 4);
if (confidence > threshold)
{
// get classes probability
Cv2.MinMaxLoc(prob.Row[i].ColRange(prefix, prob.Cols), out _, out Point max);
int classes = max.X;
var probability = prob.At <float>(i, classes + prefix);
// more accuracy
if (probability > threshold)
{
// if it is not first - increase object number, else set 1
if (objectNumbers.ContainsKey(classes))
{
objectNumbers[classes]++;
}
else
{
objectNumbers[classes] = 1;
}
// get center and width/height
var centerX = prob.At <float>(i, 0) * image.Width;
var centerY = prob.At <float>(i, 1) * image.Height;
var width = prob.At <float>(i, 2) * image.Width;
var height = prob.At <float>(i, 3) * image.Height;
objects.Add(new
{
Name = $"{_labels[classes]} #{objectNumbers[classes]}",
CenterX = centerX,
CenterY = centerY,
Width = width,
Height = height,
Probability = probability * 100
});
// label formatting
var label = $"{_labels[classes]} #{objectNumbers[classes]}";
result.Add($"{label} {probability * 100:0.00}%");
// avoid left side over edge
var x1 = (centerX - width / 2) < 0 ? 0 : centerX - width / 2;
// draw result
image.Rectangle(
new Point(x1, centerY - height / 2),
new Point(centerX + width / 2, centerY + height / 2), _colors[classes],
2
);
var textSize = Cv2.GetTextSize(label, HersheyFonts.HersheyTriplex, 0.5, 1,
out var baseline);
Cv2.Rectangle(
image,
new Rect(
new Point(x1, centerY - height / 2 - textSize.Height - baseline),
new Size(textSize.Width, textSize.Height + baseline)
),
_colors[classes],
Cv2.FILLED
);
Cv2.PutText(
image,
label,
new Point(x1, centerY - height / 2 - baseline),
HersheyFonts.Italic,
0.5,
UiServices.GetReadableForeColor(_colors[classes])
);
}
}
}
var json = JsonConvert.SerializeObject(objects, Formatting.Indented);
return new Tuple <Mat, List <string>, long, string>(image, result, time, json);
}));
}
public static BindableCollection <YoloModel> GetYoloModels()
{
const string cocoPath = "dataset/coco";
const string vocPath = "dataset/voc";
const string tiny = "tiny";
try
{
var cocoNames = File.ReadAllLines($"{cocoPath}.names");
var vocNames = File.ReadAllLines($"{vocPath}.names");
var cocoColors = Enumerable.Repeat(false, cocoNames.Length).Select(x => Scalar.RandomColor()).ToArray();
var vocColors = Enumerable.Repeat(false, vocNames.Length).Select(x => Scalar.RandomColor()).ToArray();
// weights
// coco.weights https://pjreddie.com/media/files/yolov2.weights
// cocotiny.weights https://pjreddie.com/media/files/yolov2-tiny.weights
// voc.weights https://pjreddie.com/media/files/yolov2-voc.weights
// voctiny.weights https://pjreddie.com/media/files/yolov2-tiny-voc.weights
var yoloModels = new BindableCollection <YoloModel>
{
new YoloModel
{
Name = "Coco",
_cfg = $"{cocoPath}.cfg",
_labels = cocoNames,
_colors = cocoColors,
_net = CvDnn.ReadNetFromDarknet($"{cocoPath}.cfg", $"{cocoPath}.weights")
},
new YoloModel
{
Name = "Tiny Coco",
_cfg = $"{cocoPath}{tiny}.cfg",
_labels = cocoNames,
_colors = cocoColors,
_net = CvDnn.ReadNetFromDarknet($"{cocoPath}{tiny}.cfg", $"{cocoPath}{tiny}.weights")
},
new YoloModel
{
Name = "Voc",
_cfg = $"{vocPath}.cfg",
_labels = vocNames,
_colors = vocColors,
_net = CvDnn.ReadNetFromDarknet($"{vocPath}.cfg", $"{vocPath}.weights")
},
new YoloModel
{
Name = "Tiny Voc",
_cfg = $"{vocPath}{tiny}.cfg",
_labels = vocNames,
_colors = vocColors,
_net = CvDnn.ReadNetFromDarknet($"{vocPath}{tiny}.cfg", $"{vocPath}{tiny}.weights")
}
};
return(yoloModels);
}
catch (Exception ex)
{
UiServices.ShowError(ex);
System.Windows.Application.Current.Shutdown();
}
return(null);
}