Hello all
@saumon
that's not a simple question, I'm afraid. Of course it depends from where I understood your question, and I don't mean to be sarcastic. Indeed I'm not a native English speaker, exactly like you.
Let's start from the beginning of your post where you say that you need your program performing some kind of detection/recognition of form(s) in an image, most probably a photo. Here the answers to your question are mainly two or three and they depends all from the kind of object(form) you need extract from the image. If your object is a geometric shape , let's say a circle/ellipse or a rectangle, or it is made in some way by straight( or nearly ) lines, your main and easiest option is to learn Hugh transform for line and circle detection (not hard). If the object is always geometric in nature and has a well distinct colour which, once you have applied a threshold filter to the image, forms a blob distinct from the rest of the image's elements, you can simply use Blobscanner for detecting this blob, and from there manipulating the correspondent pixels in the original image.
If instead you want to detect and manipulate a complex, not static and irregular shape, well things get quite nasty. But if you are interested to what it's inside the shape, without caring too much about the shape itself, for example you want detect the eyes in an image of a face, or the head of a chicken in an image of a hen-house, things ease a little as you have quite few algorithm to choose from that aren't that hard, like template matching or features detection/extraction like SURF or SIFT. You can find an implementation of SURF in the contributed library on
http://processing.org/reference/libraries/#computer_vision. My advice, it is to learn template matching since this is not very difficult to implement.
I search but i don't find anything which talk about picture modification with processing
I think you don't want just modify a picture, but you need a complex system to manipulate image, and for this again you should look in the library page on processing.org, under computer vision to see what it's already available. Also to put together some thing like that, if you are willing to learn few things, my personal advice here is to start searching google for shape detection, feature detection and extraction, Hugh transform, pattern recognition, edge detection, corners detection, blob detection, template matching and so on, trying to learn as much as you can on these arguments. But obviously this depends entirely by your needs.
Good luck
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
"If you would know what the Lord God thinks of money, you have only to look at those to whom he gives it."
(Maurice Baring)
