public void CreateDistortedImage()
{
var im = "trainingdata/documents/1040/other/tax1040.gif";
var f = new ImageView((Bitmap)Image.FromFile(im), ImageDistorter.DistortImage((Bitmap)Image.FromFile(im)));
f.Show();
Application.Run(f);
}
//Take categroized data and generate random distortions on the dataset. This will generate 10 distorted images per input image
public void GenerateTrainingData()
{
//create a new training set
var ts = new List <TrainingSet>();
//Create a root node with lable documents and give it two children, 1040 and 990
var root = new Category("documents");
var doc_1040 = root.AddSubcategory("1040");
var doc_990 = root.AddSubcategory("990");
//further define tree structure
var doc_1040_2012 = doc_1040.AddSubcategory("2012");
var doc_1040_2010 = doc_1040.AddSubcategory("2010");
var doc_1040_other = doc_1040.AddSubcategory("other");
//add data to the tree
ts.Add(new TrainingSet(path + "/trainingdata/documents/1040/other/tax1040.gif", doc_1040_other));
ts.Add(new TrainingSet(path + "/trainingdata/documents/1040/other/tax1040_2.png", doc_1040_other));
ts.Add(new TrainingSet(path + "/trainingdata/documents/1040/2012/tax1040_3.gif", doc_1040_2012));
ts.Add(new TrainingSet(path + "/trainingdata/documents/1040/2010/tax1040_4.jpg", doc_1040_2010));
ts.Add(new TrainingSet(path + "/trainingdata/documents/1040/other/tax1040_7.gif", doc_1040_other));
ts.Add(new TrainingSet(path + "/trainingdata/documents/1040/2010/tax1040_9.jpg", doc_1040_2010));
ts.Add(new TrainingSet(path + "/trainingdata/documents/1040/2012/tax1040_10.gif", doc_1040_2012));
ts.Add(new TrainingSet(path + "/trainingdata/documents/990/tax990.jpg", doc_990));
ts.Add(new TrainingSet(path + "/trainingdata/documents/990/tax990_2.jpg", doc_990));
ts.Add(new TrainingSet(path + "/trainingdata/documents/990/tax990_3.jpg", doc_990));
ts.Add(new TrainingSet(path + "/trainingdata/documents/990/tax990_4.jpg", doc_990));
ts.Add(new TrainingSet(path + "/trainingdata/documents/990/tax990_5.jpg", doc_990));
ts.Add(new TrainingSet(path + "/trainingdata/documents/990/tax990_6.jpg", doc_990));
int count = 0;
//loop through the tree and generate 10 distorted vesrions of every image
foreach (var td in ts)
{
var name = path + "/generateddata";
var cat = td.Category;
var paths = new List <string>();
if (!Directory.Exists(name))
{
Directory.CreateDirectory(name);
}
while (cat != null)
{
paths.Add("/" + cat.Name);
cat = cat.Parent;
}
paths.Reverse();
foreach (var p in paths)
{
name += p;
if (!Directory.Exists(name))
{
Directory.CreateDirectory(name);
}
}
for (int i = 0; i < 10; i++)
{
//Distort the image and save it to disk
var im = ImageDistorter.DistortImage((Bitmap)Image.FromFile(td.ImageFile));
using (MemoryStream memory = new MemoryStream())
{
using (FileStream fs = new FileStream(name + "/" + count + ".jpg", FileMode.Create, FileAccess.ReadWrite))
{
im.Save(memory, ImageFormat.Jpeg);
byte[] bytes = memory.ToArray();
fs.Write(bytes, 0, bytes.Length);
}
}
count++;
}
}
}
public void GenerateTrainingData()
{
var ts = new List <TrainingSet>();
var root = new Category("documents");
var doc_1040 = root.AddSubcategory("1040");
var doc_990 = root.AddSubcategory("990");
var doc_1040_2012 = doc_1040.AddSubcategory("2012");
var doc_1040_2010 = doc_1040.AddSubcategory("2010");
var doc_1040_other = doc_1040.AddSubcategory("other");
ts.Add(new TrainingSet("trainingdata/documents/1040/other/tax1040.gif", doc_1040_other));
ts.Add(new TrainingSet("trainingdata/documents/1040/other/tax1040_2.png", doc_1040_other));
ts.Add(new TrainingSet("trainingdata/documents/1040/2012/tax1040_3.gif", doc_1040_2012));
ts.Add(new TrainingSet("trainingdata/documents/1040/2010/tax1040_4.jpg", doc_1040_2010));
ts.Add(new TrainingSet("trainingdata/documents/1040/other/tax1040_7.gif", doc_1040_other));
ts.Add(new TrainingSet("trainingdata/documents/1040/2010/tax1040_9.jpg", doc_1040_2010));
ts.Add(new TrainingSet("trainingdata/documents/1040/2012/tax1040_10.gif", doc_1040_2012));
ts.Add(new TrainingSet("trainingdata/documents/990/tax990.jpg", doc_990));
ts.Add(new TrainingSet("trainingdata/documents/990/tax990_2.jpg", doc_990));
ts.Add(new TrainingSet("trainingdata/documents/990/tax990_3.jpg", doc_990));
ts.Add(new TrainingSet("trainingdata/documents/990/tax990_4.jpg", doc_990));
ts.Add(new TrainingSet("trainingdata/documents/990/tax990_5.jpg", doc_990));
ts.Add(new TrainingSet("trainingdata/documents/990/tax990_6.jpg", doc_990));
int count = 0;
foreach (var td in ts)
{
var name = "generateddata";
var cat = td.Category;
var paths = new List <string>();
if (!Directory.Exists(name))
{
Directory.CreateDirectory(name);
}
while (cat != null)
{
paths.Add("/" + cat.Name);
cat = cat.Parent;
}
paths.Reverse();
foreach (var p in paths)
{
name += p;
if (!Directory.Exists(name))
{
Directory.CreateDirectory(name);
}
}
for (int i = 0; i < 10; i++)
{
var im = ImageDistorter.DistortImage((Bitmap)Image.FromFile(td.ImageFile));
using (MemoryStream memory = new MemoryStream())
{
using (FileStream fs = new FileStream(name + "/" + count + ".jpg", FileMode.Create, FileAccess.ReadWrite))
{
im.Save(memory, ImageFormat.Jpeg);
byte[] bytes = memory.ToArray();
fs.Write(bytes, 0, bytes.Length);
}
}
count++;
}
}
}
