My application detects a foreign object (blob, cluster etc) in the live webcam image and displays object's outline on top of the image. To achieve that I employ Image and Canvas elements as follows:
<Border x:Name="ViewportBorder" Grid.Column="0" Grid.Row="1" BorderThickness="3" Background="AliceBlue" BorderBrush="Red">
<Grid>
<Image x:Name="videoPlayer" Stretch="Uniform" MouseDown="videoPlayer_MouseDown"></Image>
<Canvas x:Name="ObjectsCanvas"></Canvas>
</Grid>
</Border>
Border element in the above XAML is used just to draw a thick red line border around the Grid containing videoPlayer and ObjectsCanvas. Stretch="Uniform" is set to preserve image aspect ratio while being able it to stretch when application window gets maximized.
Every time the new frame arrives from the camera videoPlayer.Source gets updated with frame's bitmap whereas blob detection method yields a list of coordinates used for drawing a Polyline. The Polyline object is then added to ObjectsCanvas to be shown on top of the actual image frame.
Here's a part which draws the blob and adds it to the ObjectsCanvas.Children:
private void DrawBlob(List<Point> corners)
{
this.Dispatcher.Invoke(() =>
{
var myPolyline = new Polyline();
myPolyline.Stroke = System.Windows.Media.Brushes.Yellow;
myPolyline.StrokeThickness = 4;
myPolyline.FillRule = FillRule.EvenOdd;
myPolyline.Points = corners;
Canvas.SetLeft(myPolyline, 0);
Canvas.SetTop(myPolyline, 0);
ObjectsCanvas.Children.Clear(); // remove any old blob polyline
ObjectsCanvas.Children.Add(myPolyline); // add new polyline
});
}
When running the application I observe imperfect overlap of the blob object (thick yellow polyline), it gets somewhat right-shifted as shown in the image below.
Observed imperfection is not due to blob detection algorithm! I verified that by drawing the polylines of very same coordinates using old-fashion GDI methods on the actual bitmap.
It gets worse when I maximize the application window, an action causing videoPlayer to stretch:
I tried setting HorizontalAlignment and VerticalAlignment properties of ObjectsCanvas to Stretch but that does not help. Is there any method to align canvas exactly with the actual displayed image region?
I could get back to drawing the polylines using GDI, but I think it's a shame doing so in WPF...
I think I found a solution.
So, in order to stretch your canvas up to your image size you could wrap your canvas in ViewvBox control and bind your Canvas.Height and Canvas.Width to the image source's Height and Width like so:
<Grid>
<Image x:Name="MyFrame" Stretch="Uniform" />
<Viewbox Stretch="Uniform">
<Canvas
x:Name="MyCanvas"
Width="{Binding ElementName=MyFrame, Path=Source.Width}"
Height="{Binding ElementName=MyFrame, Path=Source.Height}">
<Canvas.Background>
<SolidColorBrush Opacity="0" Color="White" />
</Canvas.Background>
</Canvas>
</Viewbox>
</Grid>
However you need to set your Canvas.Background (you can make it transparent like in the example above), or ViewvBox will hide all of the canvas children otherwise.
Here are some screenshots of it working with a yellow polyline:
One more thing to note here - the coordinates of your polyline should be relative to image resolution.
Hope that helps!
I have an WPF usercontrol which is used in a winforms applications. WPF usercontrol is contained within an ElementHost container.
This WPF usercontrol has some images, button with images, labels, etc.
I have a dictionary with some geometries, the following is one of them.
<Geometry x:Key="hlpGeometry">F0 M22,22z M0,0z M11,22C17.0751,22 22,17.0751 22,11 22,4.92487 17.0751,0 11,0 4.92487,0 0,4.92487 0,11 0,17.0751 4.92487,22 11,22z M12.1901,16.6889L10.2662,16.6889 10.2662,18.6998 12.1901,18.6998 12.1901,16.6889z M8.18758,5.59005C7.40125,6.43439,7.00808,7.55265,7.00808,8.94484L8.72898,8.94484C8.76121,8.10695 8.89334,7.46564 9.12537,7.02091 9.53787,6.2217 10.2823,5.82209 11.3587,5.82209 12.2288,5.82209 12.8508,6.05412 13.2246,6.51818 13.6049,6.98224 13.795,7.53009 13.795,8.16173 13.795,8.61291 13.6661,9.04797 13.4083,9.46691 13.2665,9.70539 13.0796,9.9342 12.8475,10.1533L12.0741,10.9171C11.3329,11.6454 10.8527,12.2932 10.6336,12.8604 10.4144,13.4211 10.3049,14.1623 10.3049,15.084L12.0258,15.084C12.0258,14.2719 12.116,13.6596 12.2965,13.2471 12.4834,12.8281 12.8862,12.319 13.505,11.7195 14.3557,10.8945 14.9197,10.2694 15.1969,9.84396 15.4804,9.41857 15.6222,8.86427 15.6222,8.18107 15.6222,7.05314 15.2387,6.12824 14.4718,5.40636 13.7112,4.67804 12.6961,4.31388 11.4263,4.31388 10.0535,4.31388 8.9739,4.73927 8.18758,5.59005z</Geometry>
<DrawingGroup x:Key="hlpDrawingGroup" ClipGeometry="M0,0 V22 H22 V0 H0 Z">
<GeometryDrawing Brush="#FF00AA2B" Geometry="{StaticResource hlpGeometry}" />
</DrawingGroup>
<DrawingImage x:Key="ico_helpDrawingImage" Drawing="{StaticResource hlpDrawingGroup}" />
I have an WPF Image and I bound above DrawingImage to it using the Source attribute. I bind the source attribute to a property in the view model.
Something like below:
<Image x:Name="MyImage"
Height="24"
Width="24"
VerticalAlignment="Center"
Source="{Binding Path=MyIcon}"/>
It is working fine when windows is scaled to 100% but when scaled to a higher one, let's say, 125%, then the image gets fuzzy.
Also the image look like gets bigger than 24x24 and it is being cut-off when I set a scale greater than 100% (125%).
How can I make image to not get fuzzy and set image to always be the same size 24x24?
You can use the UseLayoutRounding and SnapsToDevicePixels properties.
Set the UseLayoutRounding property to true to prevent blurring caused by anti-aliasing and to tell the layout system to align elements with pixel boundaries.
RenderOptions.BitmapScalingMode is mainly for larger images to be displayed more smoothly, you can choose HighQuality.
Use high quality bitmap scaling, which is slower than LowQuality mode, but produces higher quality output.
You can check this link:
https://learn.microsoft.com/en-us/dotnet/api/system.windows.media.bitmapscalingmode?redirectedfrom=MSDN&view=windowsdesktop-7.0
Codeļ¼
<Image x:Name="img" MouseWheel="img_MouseWheel"
Height="24"
Width="24"
UseLayoutRounding="True"
SnapsToDevicePixels="True"
VerticalAlignment="Center"
RenderOptions.BitmapScalingMode="HighQuality"
Source="2.jpg"/>
With images I have always found the below to always clear up the fuzzy stuff.
<Image x:Name="MyImage"
Height="24"
Width="24"
RenderOptions.BitmapScalingMode="HighQuality"
VerticalAlignment="Center"
Source="{Binding Path=MyIcon}"/>
How can I tint/colorize an image in WPF (using MVVM) without sacrificing performance? A purely XAML solution would be ideal, as modifying bitmaps in the code will cause performance loss with lots of changing images. The image is made up of more than simple shapes, so it is not possible using a path.
Unlike WinForms/GDI+, WPF does not seem to contain any easy ways to colorize/tint an image as it is being rendered. Two ideas for accomplishing this are, using a shader, or overlaying a colored rectangle over the image.
I decided to try the rectangle route and found that it works. Basically, all you need to do is overlay a colored rectangle over your image, and use an OpacityMask to restrict the color fill to a certain area. OpacityMask is primarily used with paths, but it can take any kind of brush, including an ImageBrush. This means you can use your image as a "stencil" for the colored fill.
Example: (Taken from my application where a user can "highlight" a section of a map, the actual image looks like this)
Before Overlay & Mask
After Overlay & Mask
Here is all of the required XAML for this:
<Image
Source="{Binding MyImage}"
Width="150"
Height="150" />
<Rectangle Width="150" Height="150">
<Rectangle.Fill>
<SolidColorBrush Color="{Binding Color}"/>
</Rectangle.Fill>
<Rectangle.OpacityMask>
<ImageBrush ImageSource="{Binding MyImage}"/>
</Rectangle.OpacityMask>
</Rectangle>
To bind the color to a brush as I did, use a ColorToBrushConverter.
I have clipped Image:
<Image Name="Img" Source="/UntitledImage.jpg">
<Image.Clip>
<EllipseGeometry Center="115,115" RadiusX="50" RadiusY="50"></EllipseGeometry>
</Image.Clip>
</Image>
or:
<Image Name="oldImg" Source="/UntitledImage.jpg">
<Image.Clip>
<RectangleGeometry Rect="115,115,50,50"></RectangleGeometry>
</Image.Clip>
</Image>
I want added blur for each edge for Image after clip.
I want manage thick blur area for Image.
Is it possible?
Rather than creating a blurred version of each image, why not put a semi-transparent image over the top of the image to give the appearance of a blurred edge instead?
This would, I expect, be much quicker and simpler.
This solution might be a little heavier for your purpose, but I suppose it will serve any need you could have.
You can try to use Nokia Imaging SDK to process your images. The SDK usage is very simple and well documented In the developer library.
Also, before you decide, the link above contains sample apps to explore the effects.
I am arranging quite freely a lot of elements on a Canvas layout, in fact the elements represent an interactive flow-chart. As transformations are applied, I need relative transformations on some of the elements.
Especially I require some elements being anchored or docked to their parent elements. I found different solutions, however I don't know if they solve my problem in the most elegant way.
Here is an example:
<Line X1="80" X2="800" Y1="730" Y2="730"/>
<Polygon Points="0,30 40,0 40,60" Canvas.Left="48" Canvas.Top="700"/>
The Polygon draws a triangle and I would like to let it dock on the left side of line. Which means, when translating the line to a new position or when scaling it down, the Polygon should move with it.
Is this possible?
put them in a canvas of their own, that way you can position the outer canvas absolutely and keep the inner stuff together.
like this:
<Canvas>
<Line X1="60" X2="820" Y1="60" Y2="760"> <!--some other line--> </Line>
<Canvas Canvas.Left="48" Canvas.Top="700">
<Polygon Points="0,30 40,0 40,60"/>
<Line X1="32" X2="752" Y1="30" Y2="30"/>
</Canvas>
</Canvas>