| Groups | Blog | Home |
dotnet drawing api : Problems with the graphics class
I'm trying to write some scientific software for my job at a university and I'm having some problems with the graphics class. I've included the source that I've used to test things. Source1.tif is a 512x512 greyscale image and source2.tif is a 3072x2304 greyscale image created by an electron microscope. In gfx1 the main problem is that the color fidelity is compromised somehow as the the color in the original image is FF and in the output image FE and so on for every pixel. I've tried to just clone the bitmap image and the result is the same. The other problem is why do I have to use the pixeloffsetmode.half? Without it 2 pixels on the top and left edge gets chopped off and to the right and bottom I get and transparent edge. In gfx2 I'm trying to copy an bitmap to a larger one but it always resizes it no matter what I do. Is there a way to get pixel->pixel copy without any distortions? Is there a way to save just a part of a bitmap/graphics without copying it first? Can anyone recommend any good books covering graphics/bitmaps in the NET Framework? Thanks for any help! Linus Dim source1 As Bitmap = New Bitmap("K:\source1.tif") Dim source2 As Bitmap = New Bitmap("K:\source2.tif") Dim dest1 As Bitmap = New Bitmap(1024, 1024) Dim dest2 As Bitmap = New Bitmap(10000, 2304) Dim gfx1 As Graphics = Graphics.FromImage(dest1) gfx1.SmoothingMode = System.Drawing.Drawing2D.SmoothingMode.None gfx1.InterpolationMode = Drawing2D.InterpolationMode.NearestNeighbor gfx1.PixelOffsetMode = Drawing2D.PixelOffsetMode.Half gfx1.CompositingQuality = Drawing2D.CompositingQuality.AssumeLinear gfx1.CompositingMode = Drawing2D.CompositingMode.SourceCopy gfx1.DrawImage(source1, 0, 0, 1024, 1024) Dim gfx2 As Graphics = Graphics.FromImage(dest2) gfx2.Clear(Color.White) gfx2.SmoothingMode = System.Drawing.Drawing2D.SmoothingMode.None gfx2.InterpolationMode = Drawing2D.InterpolationMode.NearestNeighbor gfx2.PixelOffsetMode = Drawing2D.PixelOffsetMode.Half gfx2.CompositingQuality = Drawing2D.CompositingQuality.AssumeLinear gfx2.CompositingMode = Drawing2D.CompositingMode.SourceCopy gfx2.PageUnit = GraphicsUnit.Pixel gfx2.DrawImageUnscaled(source2, 0, 0) gfx2.DrawImageUnscaledAndClipped(source2, New Rectangle(3024, 0, 10000, 2304)) dest1.Save("k:\dest1.png", System.Drawing.Imaging.ImageFormat.Png) dest2.Save("k:\dest2.png", System.Drawing.Imaging.ImageFormat.Png)
Whenever you resize an image the Graphics object needs to reposition the
pixels in a given ratio. The values must eventually be converted to intefers at some time so rounding losses mean that images can move by a pixel or two. The PixelOffset.Half mode effectively adds a float value of 0.5 to the calculation such that the integer rounding goes to the nearest value and is not automatically rounded down. Your bitmaps are resizing because you're not taking the declared resolutions of the originals and the default resolution of a new bitmap into account. When you create a new image you should set the resolution from the old one or use the DrawImage method that takes specific source and destination rectangles and uses the "Pixel" graphics unit. -- -- Bob Powell [MVP] Visual C#, System.Drawing Ramuseco Limited .NET consulting http://www.ramuseco.com Find great Windows Forms articles in Windows Forms Tips and Tricks http://www.bobpowell.net/tipstricks.htm Answer those GDI+ questions with the GDI+ FAQ http://www.bobpowell.net/faqmain.htm All new articles provide code in C# and VB.NET. Subscribe to the RSS feeds provided and never miss a new article. [quoted text, click to view] "Linus" <lsilvand@hotmail.com> wrote in message
news:1190889357.524434.123620@d55g2000hsg.googlegroups.com... > Hi, > > I'm trying to write some scientific software for my job at a > university and I'm having some problems with the graphics class. I've > included the source that I've used to test things. > Source1.tif is a 512x512 greyscale image and source2.tif is a > 3072x2304 greyscale image created by an electron microscope. In gfx1 > the main problem is that the color fidelity is compromised somehow as > the the color in the original image is FF and in the output image FE > and so on for every pixel. I've tried to just clone the bitmap image > and the result is the same. The other problem is why do I have to use > the pixeloffsetmode.half? Without it 2 pixels on the top and left edge > gets chopped off and to the right and bottom I get and transparent > edge. > In gfx2 I'm trying to copy an bitmap to a larger one but it always > resizes it no matter what I do. Is there a way to get pixel->pixel > copy without any distortions? > Is there a way to save just a part of a bitmap/graphics without > copying it first? > Can anyone recommend any good books covering graphics/bitmaps in the > NET Framework? > > Thanks for any help! > Linus > > Dim source1 As Bitmap = New Bitmap("K:\source1.tif") > Dim source2 As Bitmap = New Bitmap("K:\source2.tif") > Dim dest1 As Bitmap = New Bitmap(1024, 1024) > Dim dest2 As Bitmap = New Bitmap(10000, 2304) > > Dim gfx1 As Graphics = Graphics.FromImage(dest1) > gfx1.SmoothingMode = > System.Drawing.Drawing2D.SmoothingMode.None > gfx1.InterpolationMode = > Drawing2D.InterpolationMode.NearestNeighbor > gfx1.PixelOffsetMode = Drawing2D.PixelOffsetMode.Half > gfx1.CompositingQuality = > Drawing2D.CompositingQuality.AssumeLinear > gfx1.CompositingMode = Drawing2D.CompositingMode.SourceCopy > gfx1.DrawImage(source1, 0, 0, 1024, 1024) > > Dim gfx2 As Graphics = Graphics.FromImage(dest2) > gfx2.Clear(Color.White) > gfx2.SmoothingMode = > System.Drawing.Drawing2D.SmoothingMode.None > gfx2.InterpolationMode = > Drawing2D.InterpolationMode.NearestNeighbor > gfx2.PixelOffsetMode = Drawing2D.PixelOffsetMode.Half > gfx2.CompositingQuality = > Drawing2D.CompositingQuality.AssumeLinear > gfx2.CompositingMode = Drawing2D.CompositingMode.SourceCopy > gfx2.PageUnit = GraphicsUnit.Pixel > gfx2.DrawImageUnscaled(source2, 0, 0) > gfx2.DrawImageUnscaledAndClipped(source2, New Rectangle(3024, > 0, 10000, 2304)) > > dest1.Save("k:\dest1.png", > System.Drawing.Imaging.ImageFormat.Png) > dest2.Save("k:\dest2.png", > System.Drawing.Imaging.ImageFormat.Png) >
On Sep 29, 2:26 pm, "Bob Powell [MVP]" <b...@spamkillerbobpowell.net>
[quoted text, click to view] wrote: > Whenever you resize an image theGraphicsobject needs to reposition the > pixels in a given ratio. The values must eventually be converted to intefers > at some time so rounding losses mean that images can move by a pixel or two. > The PixelOffset.Half mode effectively adds a float value of 0.5 to the > calculation such that the integer rounding goes to the nearest value and is > not automatically rounded down. > > Your bitmaps are resizing because you're not taking the declared resolutions > of the originals and the default resolution of a new bitmap into account. > When you create a new image you should set the resolution from the old one > or use the DrawImage method that takes specific source and destination > rectangles and uses the "Pixel"graphicsunit. > So do I understand it correctly that the graphics object being manipulated is always being manipulated in single/double and not integers? And if I use PixelOffset it might shift the image too far in the other direction? I'm trying to use the DrawImage method using the upperleftcorner copy with the pixel graphicsunit and it still resizes it. The DrawImage method is a bit confusing. Are there any resources that describes the graphics class in more indepth because the MSDN doesn't do it. Or am I better off using the bitmaps and creating my own methods using getpixel and setpixel? Thanks for the help! Linus
The graphics class is very complex but at the same time well thought
out. You certainly don't need to invent your own method for such a simple case and certainly do not mess with getpixel and setpixel unless you want the slowest graphics possible. Calculations are al done in single precision floating point. this enables the transformation pipeline to work correctly. Integers would be too inaccurate. The pixel offset mode is only to lock the pixel to a specific part of a pixel, it won't move the image too far. What's happening is that the change in size is causing interpolation which is not always good. Try using the bilinear mode instead of the bicubic. Finally, for the best control, use the DrawImage(Image,Rectangle,int,int,int,int,GraphicsUnit) method. -- Bob Powell [MVP] Visual C#, System.Drawing Ramuseco Limited .NET consulting http://www.ramuseco.com Find great Windows Forms articles in Windows Forms Tips and Tricks http://www.bobpowell.net/tipstricks.htm Answer those GDI+ questions with the GDI+ FAQ http://www.bobpowell.net/faqmain.htm All new articles provide code in C# and VB.NET. Subscribe to the RSS feeds provided and never miss a new article. [quoted text, click to view] Linus wrote:
> On Sep 29, 2:26 pm, "Bob Powell [MVP]" <b...@spamkillerbobpowell.net> > wrote: >> Whenever you resize an image theGraphicsobject needs to reposition the >> pixels in a given ratio. The values must eventually be converted to intefers >> at some time so rounding losses mean that images can move by a pixel or two. >> The PixelOffset.Half mode effectively adds a float value of 0.5 to the >> calculation such that the integer rounding goes to the nearest value and is >> not automatically rounded down. >> >> Your bitmaps are resizing because you're not taking the declared resolutions >> of the originals and the default resolution of a new bitmap into account. >> When you create a new image you should set the resolution from the old one >> or use the DrawImage method that takes specific source and destination >> rectangles and uses the "Pixel"graphicsunit. >> > > So do I understand it correctly that the graphics object being > manipulated is always being manipulated in single/double and not > integers? And if I use PixelOffset it might shift the image too far in > the other direction? > I'm trying to use the DrawImage method using the upperleftcorner copy > with the pixel graphicsunit and it still resizes it. The DrawImage > method is a bit confusing. Are there any resources that describes the > graphics class in more indepth because the MSDN doesn't do it. Or am I > better off using the bitmaps and creating my own methods using > getpixel and setpixel? > > Thanks for the help! > > Linus
Thanks for taking the time to answer my questions. I still have
problems with the class. When loading(is this a bug?)/stretching a greyscale or indexed tiff image the values of the pixel gets shifted with 1 down (for example from $FF to $FE) if I use NearestNeighbor as the interpolation method. Any other method (bicubic or bilinear...) mangles the data beyond use for me. I've used the DrawImage(Image,Rectangle,int,int,int,int,GraphicsUnit) method with pixel as graphicsunit but it doesn't help. Going through a lot of different tiff images, I've noticed that Drawing2D.PixelOffsetMode.Half doesn't always help as sometimes it leaves a transparent line at the right and bottom edges. As I said earlier, I'm writing some scientific software and if I can't count on getting out non-mangled images it's of little use continuing. The only solution I see is programming my own methods with Setpixel and Getpixel but these are insanely slow! Are there any other alternatives? I've looked at the "Victor Image Processing Library" but it's a bit too expensive for me and what I'm doing. One more question: image0.setpixel(0,0,color.green) if image0.getpixel(0,0)=color.green then ...... endif Why doesn't this work? If I use if image0.getpixel(0,0).toargb=color.green.toargb then then it works. Why is this? What's the point of having to convert a structure into an integer to get it working? Thanks, Linus [quoted text, click to view] > Thegraphicsclassis very complex but at the same time well thought > out. You certainly don't need to invent your own method for such a > simple case and certainly do not mess with getpixel and setpixel unless > you want the slowestgraphicspossible. > > Calculations are al done in single precision floating point. this > enables the transformation pipeline to work correctly. Integers would be > too inaccurate. > > The pixel offset mode is only to lock the pixel to a specific part of a > pixel, it won't move the image too far. What's happening is that the > change in size is causing interpolation which is not always good. Try > using the bilinear mode instead of the bicubic. > > Finally, for the best control, use the > DrawImage(Image,Rectangle,int,int,int,int,GraphicsUnit) method. > > -- > Bob Powell [MVP] > Visual C#, System.Drawing > > Ramuseco Limited .NET consultinghttp://www.ramuseco.com > > Find great Windows Forms articles in Windows Forms Tips and Trickshttp://www.bobpowell.net/tipstricks.htm > > Answer those GDI+ questions with the GDI+ FAQhttp://www.bobpowell.net/faqmain.htm > > All new articles provide code in C# and VB.NET. > Subscribe to the RSS feeds provided and never miss a new article. > > Linus wrote: > > On Sep 29, 2:26 pm, "Bob Powell [MVP]" <b...@spamkillerbobpowell.net> > > wrote: > >> Whenever you resize an image theGraphicsobject needs to reposition the > >> pixels in a given ratio. The values must eventually be converted to intefers > >> at some time so rounding losses mean that images can move by a pixel or two. > >> The PixelOffset.Half mode effectively adds a float value of 0.5 to the > >> calculation such that the integer rounding goes to the nearest value and is > >> not automatically rounded down. > > >> Your bitmaps are resizing because you're not taking the declared resolutions > >> of the originals and the default resolution of a new bitmap into account. > >> When you create a new image you should set the resolution from the old one > >> or use the DrawImage method that takes specific source and destination > >> rectangles and uses the "Pixel"graphicsunit. > > > So do I understand it correctly that thegraphicsobject being > > manipulated is always being manipulated in single/double and not > > integers? And if I use PixelOffset it might shift the image too far in > > the other direction? > > I'm trying to use the DrawImage method using the upperleftcorner copy > > with the pixel graphicsunit and it still resizes it. The DrawImage > > method is a bit confusing. Are there any resources that describes the > >graphicsclassin more indepth because the MSDN doesn't do it. Or am I > > better off using the bitmaps and creating my own methods using > > getpixel and setpixel? > > > Thanks for the help! > > > Linus
Don't see what you're looking for? Try a search.
|
June 2003
July 2003
August 2003
September 2003
October 2003
November 2003
December 2003
January 2004
February 2004
March 2004
April 2004
May 2004
June 2004
July 2004
August 2004
September 2004
October 2004
November 2004
December 2004
January 2005
February 2005
March 2005
April 2005
May 2005
June 2005
July 2005
August 2005
September 2005
October 2005
November 2005
December 2005
January 2006
February 2006
March 2006
April 2006
May 2006
June 2006
July 2006
August 2006
September 2006
October 2006
November 2006
December 2006
January 2007
February 2007
March 2007
April 2007
May 2007
June 2007
July 2007
August 2007
September 2007
October 2007
November 2007
December 2007
January 2008
February 2008
March 2008
April 2008
May 2008
June 2008
| home · blog · groups | Questions? Comments? Contact the dn |