Develop Reference

Winforms: Capture part of PDF image to OCR

This is a C# winforms question.
The process I am trying to achieve is the following:
Using AxAcroPDFLib I'm loading a pdf file to the form
I want the user to be able to specify a square on that PDF and create a bmp from it
That bmp will then be loaded to a OCR to become text
What is my issue:
Step 1 and 3 are easy to do, the problem is how to allow the user to draw a square on top of the AxAcroPDFLib for a screenshot.
I already got different ways to draw squares on native winform components, but AxAcroPDFLib does not support mouse down, up, move, etc and paint events.
There is the option to convert the PDF to bitmap and display it on a picturebox and deal with events for drawing the square. Problem with that approach is that my PDF's are usually more than two pages, and I would like to avoid the conversion pdf to bmp due to changes to image quality that will impact on OCR.
I came to think that maybe something that works as the windows snippingtool would do the the job. My application would get the screenshot, temporarily save the image on disk (must be a file for OCR), I would then pass it to the OCR and done. Hard part, I could not think on how to take the screenshot of part of the PDF.
Do anyone here have any suggestion to different components or workarounds to deal with the requirement above? I am using Adobe just because it is simple, but maybe there are other components better suited to handle my requirements? I googled but haven't found any free ones, trying to avoid paid options.
Thanks

At some point in this process, the PDF is going to have to be rasterized in order to be passed to the OCR, so I don't totally understand your objection to converting it to a bitmap. If you're okay with Snipping Tool's behavior, then you must be OK with the quality of the PDF control's PDF->screen rasterization. If that resolution is acceptable, why not just capture the control's content to a Bitmap and let the user draw the selection marquee over that Bitmap?
Here is code I'm using to capture a control's contents to a Bitmap in Windows Forms. One caveat is that this is really a screen capture, so any windows or controls that overlap the control's visible area will be captured in the image.
using (Bitmap b = new Bitmap(width, height))
{
using (Graphics g = Graphics.FromImage(b))
{
Point p1 = myControl.PointToScreen(new Point(0, 0));
g.CopyFromScreen(p1.X, p1.Y, 0, 0, size);
}
// do stuff here with your Bitmap
}

C# Fast load Bitmap into PictureBox

I receve Bitmap image from a camera at 30 fps, and I need to display all images in a pictureBox.
The problem is that the PictureBox is very slow!
I have try to implement a custom PictureBox with DoubleBuffer enabled but the problem is not resolved.
Do you have a custom PictureBox or an user control or a solution that can display the image faster?
Additional information:
The image resolution is 2048x1088 with 256 graylevel (8bit image).
I use AForge.NET for elaborate the images.
Thank you

That image gets expensive to draw when it has to be resized to fit the PB's client area. Which is very likely in your case because your images are pretty large. It uses a high-quality bi-cubic filter to make the resized image look good. That's pretty expensive, albeit that the result is good.
To avoid that expense, resize the image yourself before assigning it to the Image property. Make it just as large as the PB's ClientSize.
That's going to make a big difference in itself. The next thing you can do is to create the scaled bitmap with the 32bppPArgb pixel format. It's the format that's about 10 times faster then any other because it matches the video adapter on most machines so no pixel format conversions are necessary.

How to WIC Encode a D2D bitmap to png

At the moment I'm programming a paint app in WPF C# and have a problem. I want to draw different shapes and save them as a png/bmp. So right now I'm decoding a bmp to a direct2d bitmap to draw on rendertarget. But my problem is that I don`t know how to save the D2DBitmap as a png because I can not find a encode function...
I am not allowed to use SharpDX so I hope someone can help me with this problem.
Or maybe somebody has another solution for drawing shapes and stuff but NOT in the xml file.

So I managed it by myself.
I created a surfaceRenderTarget that can be painted on and also is shown on Screen.
Then I created a WicBitmapRenderTarget with the same properties and this Rendertarget is easily saveable with a saveToFile method.
So I draw on the surfaceRenderTarget and push every operation to an operationstack and when I click the save button the programm draws the hole stack on the bitMapTarget and saves this.
This is my solution, maybe someone has a better.

Clear drawn graphics on a picturebox without clearing original picture

In a geographical software written in C#, a PictureBox is used to show GIS map that is saved as a png file in a temporary directory. There is some geometric shapes we need to be drawn on map. We used System.Drawing methods to perform this action.
Sometimes we need to change some properties these shapes or delete them, we need to remove the shapes without making beneath them black. Drawing them again with Color.Transparent obviously doesn't work, using Graphics#Clear(Color.Transparent) doesn't work too for the same reason.
We even tried using another picture box with transparent background that is used only for purpose of drawing shapes on; so that when we use Graphics#Clear(Color.Transparent) map container remains untouched. Sounded like a perfect idea at first, but because i don't know how and why it makes map container PictureBox invisible and map viewer panel is totally black, this idea failed too.
MapViewerForm
|-- Toolbar
|-- StatusBar
|-- MapViewer Panel (Provides scrollbars)
|-- MapContainer Pictutebox
|-- Shapes drawing canvas PictureBbox (The same size and location as map container, only difference is z-order)
I prefer to use the two PictureBoxes and making 'layers' idea, i think it's less unprofessional than the other idea (I am actually a java developer and this is a C# project!), I think there should be something like java's JLayeredPane in C# to adjust z-order of those two picture boxes and omit black screen bug, But if there is a solution to draw shapes on map container itself and then clear them without losing portions of maps lying behind them i'd appreciate that answer too.
P.S: If we load map picture from file and store it in a Bitmap or Image private field and when we need to clear drawings, load image from that field with a piece of code like picMapArea.Image = MapViewer.getInstance().getMapImage(); (Note: MapViewer is a singleton class) the painted shapes will be gone but it's obviously not anything like a "good idea" because of poor performance and lagging.
Thanks in advance.

Simply draw the shapes in an event handler for the picturebox Paint event.
To restore the view, all you have to do is call the picturebox Invalidate() method, so it repaints the Image, and not draw anything in your Paint event handler.

Just use an additional Bitmap:
Bitmap original = LoadBitmap(...);
Bitmap copy = new Bitmap(original);
Graphics graph = Graphics.FromImage(copy);
// draw some extra
PictureBox1.Image = copy;

image processing in c#

before question think about for example photoshop. When you draw a rectangle on the picture.You can move it. And when you move it works very quickly and it doeasnt make some traces on the picture.
So my question is, how to do that in c# application?

This might be useful for you
Image Processing for Dummies with C# and GDI+ Part 1 - Per Pixel Filters
Image Processing for Dummies with C# and GDI+ Part 2 - Convolution Filters
Image Processing for Dummies with C# and GDI+ Part 3 - Edge Detection Filters
Image Processing for Dummies with C# and GDI+ Part 4 - Bilinear Filters and Resizing
Image Processing for Dummies with C# and GDI+ Part 5 - Displacement filters, including swirl
Image Processing for Dummies with C# and GDI+ Part 6 - The HSL color space

When you are moving the rectangle, Photoshop doesn't put it in the image and then draw the image, instead the image is drawn without the rectangle, and the rectangle is drawn on top of that on the screen. That way when you move the rectangle it can redraw the part of the image that previously was covered by the rectangle, and draw the rectangle at the new position.

I think you're asking about selection rectangles (or other temporary shapes) on top of the document image. This effect is sometimes known as “rubber banding”, especially when drawing a line from one point to another (it stretches like a rubber band).
Traditionally, this was done by using XOR drawing -- instead of overwriting the image with the selection shape, the colors in that area are inverted. Then, to remove the selection, it suffices to invert the colors again, returning to the same original image. Today, graphics rendering is fast enough that such tricks are not usually necessary; it suffices to simply repaint that part of the window (without the rectangle).
Either way, it is important to recognize that the document image — the image the user is editing — is not the same as the window image, which is just a copy to be remade whenever necessary. In the window, the document image is drawn and then selections, guide marks, and other such controls are drawn on top of it.
I'm not familiar with C#'s GUI facilities (and I understand there is more than one GUI framework you might be using), but it's probably got the usual structure of putting many "widgets", "views", or "controls" in the window (possibly nested inside each other). You can do a straightforward selection box — though not an optimally efficient one — by just putting an appropriately sized rectangle widget (with a solid border and a transparent background) on top of an image widget. This lets your GUI framework take care of the appropriate redrawing for you and is probably a good cheap way to start.