I need to develop a software that selects a face from a photo where the background is a plain color (green, like in the movies).
Then we want to compose that selection with another background image, this part is easy with many libraries. But I don't know how can I do the selection? Can you give some links or libraries to investigate? I can do this project with any language of my choose, so examples or links in any language are welcome.
Ok, what you are trying to do is called chroma key. Like you say, it used a lot in the movies with a blue/green screen. On windows its actually pretty easy to do because its built into windows as part of GDI+ (or on C#, I think its just called the Graphics class).
I dont have any sample code handy, but the process is pretty straight forward:
With GDI+, you create a bitmap object of your foreground image (the one with the green background). Then create an ImageAttributes object. Use ImageAttribute's object's SetColorKey() method to specify a color or range of colors to use as the background color. Lastly, draw that bitmap object over the target bitmap, and GDI+ will draw it as if the background color is transparent.
There's more to it in that in code, but concept-wise thats all there is to it.
This is probably an area where it is easier to work in some other space than RGB - such as HSV.
I would also look at the OpenCV library.
Related
Let's say I have an image with a few colors.
I would like to replace programmatically a specific existing color by a new one.
(something simple, no need to support gradients, like I saw elsewhere).
E.g. I have an image showing a green circle and I want to display it as a red circle (every pixel initially defined with a given (R,G,B) is now displayed with a new (R,G,B).
Any idea of how to do that with the Apple ios SDK ? (or open source ...)
And btw what would be the best image file format to make this easier (png, jpg ....) ?
Thanks !
You should be able to do this using Core Image filters. the Color Cube CI filter lets you map a source color range to destination colors. You should be able to define a source color range and map it to different colors.
That's one CI Filter I didn't figure out how to use however. If you do a search on "Color Cube" in the Xcode help system there is sample code that does a "chromakey" effect that knocks out green shades to transparent. You should be able to adapt that to your needs.
I have a project on Github called CIFilterTest that shows how to use Core Image filters to process images. It's written as a general-purpose system that lets you try a wide variety of filters that use a standard set of parameters (points, colors, 1 or 2 source images, and floating-point values.) I never did take the time to generate the 3D color mapping "cube" that the color cube filter needs as input, so it doesn't allow you to use that particular filter. You'll have to look at the color Cube sample code in the Xcode docs to generate inputs for the Color Cube filter, but my sample app should help a great deal with the basic setup for doing CI based image processing.
answered similar question here:
Replace particular color of image in iOS
in short: I would suggest using CoreImage filter.
I am writing a little application in classic asp (don't judge me, I don't have a choice) where I need to offer users a choice of PNG and a choice of colors, then I need to display the image among other html elements. The color will essentially come from a color selector, so there will be endless choices and I don't really have the option of making every PNG in every possible color. Does anybody know of a script I could use for this, or should I just take a crack at making one from scratch?
You can achieve this with ImageMagick or GraphicsMagick.
They both have a COM component you can install and use from ASP. They allow image manipulation from code.
Using .NET like suggested above will do the same thing, but it's equally possible in classic ASP (or other languages like perl or Python) using one of these two components.
Mind you: It will not be a simple thing to do, but essentially what you want is to manipulate images with code, and more specifically change one color, which is tricky in most programming languages.
Here are a few links to get you started:
Using ImageMagick from classic ASP
Replacing a color with ImageMagick
Even we can take color code from color picker and assign it as the background color something like this
color = "#3B0B0B"
I want to replace the particular color of an image with other user selected color. While replacing color of image, I want to maintain the gradient effect of that original color. for example see the attached images.
I have tried to do so with CoreGraphics & I got success to replace color. But the replacing color do not maintain the gradient effect of the original color in the image.
Can someone help me on this? Is the CoreGraphics is right way to do this?
Thanks in advance.
After some struggling almost with the same problem (but with NSImage), made a category for replacing colors in NSImage which uses ColorCube CIFilter.
https://github.com/braginets/NSImage-replace-color
inspired by this code for UIImage (also uses CIColorCube):
https://github.com/vhbit/ColorCubeSample
I do a lot of color transfer/blend/replacement/swapping between images in my projects and have found the following publications very useful, both by Erik Reinhard:
Color Transfer Between Images
Real-Time Color Blending of Rendered and Captured Video
Unfortunately I can't post any source code (or images) right now because the results are being submitted to an upcoming conference, but I have implemented variations of the above algorithms with very pleasing results. I'm sure with some tweaks (and a bit of patience) you might be able to get what you're after!
EDIT:
Furthermore, the real challenge will lie in separating the different picture elements (e.g. isolating the wall). This is not unlike Photoshop's magic wand tool which obviously requires a lot of processing power and complex algorithms (and is still not perfect).
My program needs to output a (fairly complex) form to the printer, including several images. I’m currently doing this using Delphi (2006)’s Printer.Canvas, after selecting a PDF printer (PDF995). This works like a treat.
However, I’m now running into a problem: there’s one partially transparent image that needs to be placed on top of other elements (borders, background and such), with portions of that text still visible through parts of the image.
Doing this on a regular screen Canvas works fine with regular TBitmaps, by using the TransparentColor property. However, when I try to do this on a printer, it doesn’t always work; and when I try this on a PDF printer, it never works: the background turns black, or (the best result so far), turns white, but still overwrites anything underneath it.
I’ve tried achieving the same result by inserting a PNG image with alpha transparency (a.k.a. translucency) in a Word document, and then sending that to said PDF printer. The translucency disappears, but pixel transparency is maintained. So that, at least, should somehow be possible.
Anybody know how?
You mention you are using the TransparentColor property, so is it correct to say you don't really need alpha transparency and can get away with using binary transparency (transparency for each pixel is either off or on)?
If so, it might be possible to generate a Region from your bitmap.
You then use this region as a clipping region and draw the bitmap.
The PDF printer might be able to handle a clipping region correctly.
Here's an example of generating a region from a bitmap, it's used to make non-rectangular forms but the idea is the same:
http://www.delphi-central.com/BitmapShapedForm.aspx
Here is another example of setting a clipping region for your TCanvas:
http://www.efg2.com/Lab/OtherProjects/PrinterDemo2.htm
I don't think you can achieve partial transparency using a pdf printer. What you can do is get the VisPDF package. With it you can add a mask to the images contained in the pdf that you actually produce right in you're application. If this is not an option, you could combine all your background stuff in to one image. Drawing the alphachanneled thing on this image.
I also had this problem before, If I recall correctly, what I ended up doing was to create an off screen bitmap that I manipulated, then after I was done copied again and used the new bitmap copy on the canvas I was sending to the printer.
You may want to consider using a third-party PDF component in your application to render the canvas directly to PDF. I use this technique for specialized reporting (RTF, Radioemtric JPEGs, Tables, Text etc) in a commercial product I developed for Infrared Thermography. I am very happy with the performance and quality. The component I am using llPDFLib was just updated and I believe now supports full image transparency.
Best of luck. Printing images with an Alpha channel can be challenging at times in my experience.
Is there a good image watermark control for Delphi? It would be nice if it could receive both text and images to insert in a base photo.
I would also prefer it was free.
I couldn't find any pre-packaged controls. But watermarking is not very hard at all. All you simply need to do is draw an image on top of another image with the use of alpha blending. This site has a whole section on alpha blending in Delphi. They provide links to graphics libraries which have implemented it.
However if you're using Delphi.NET, and can access the relevant classes in the framework, there is an easier way using only framework methods.
Take a look at Graphics32 together with GraphicEX.
Or see if PascalMagick does the trick.