jellyz / simple-iphone-image-processing Goto Github PK

Use the Image::createImage(UIImage*,int width,int height) method to create a 
new Image object
Notice that upon destruction of the object the m_imageData variable does not 
get released

To fix this add:
m_ownsData=true;
To the bottom of the constructor mentioned at the top of this issue.

What steps will reproduce the problem?
1. invoke getPercentSet, linking fails 

What is the expected output? What do you see instead?
I'm expecting all of the functions in Image.h to be implemented in Image.mm

What version of the product are you using? On what operating system?
svn revision 16

Please provide any additional information below.
I'm completely new to C/Obj-C/XCode/iPhone, so this may be my issue.  I see 
there are a handful of functions not implemented in Image.mm including 
getPercentSet.

What steps will reproduce the problem?
1. Improve gaussianBlur() or take my improved code
2.
3.

What is the expected output? What do you see instead?
gaussianBlur() reduces image by 5 pixels in the upper left corner.

What version of the product are you using? On what operating system?
xCode and the newest version (13th july 2010)

Please provide any additional information below.
Fixed gaussianBlur() function:

ImageWrapper *Image::gaussianBlur() {
    Image *result=new Image(m_width, m_height);
    for(int y = 0; y < m_height; y++) {
        for(int x = 0; x < m_width; x++) {

            int val = (*this)[y][x]*41; //Declare currents pixels value
            int middle = 41; //Declare average

            if (x > 0)                          { val += (*this)[y][x-1]    *26;middle += 26;}
            if (x > 1)                          { val += (*this)[y][x-2]    *7; middle += 7;}
            if (x < m_width-1)                  { val += (*this)[y][x+1]    *26;middle += 26;}
            if (x < m_width-2)                  { val += (*this)[y][x+2]    *7; middle += 7;}

            if (y > 0)                          { val += (*this)[y-1][x]    *26;middle += 26;}
            if (y > 0 && x > 0)                 { val += (*this)[y-1][x-1]  *16;middle += 16;}
            if (y > 0 && x > 1)                 { val += (*this)[y-1][x-2]  *4; middle += 4;}
            if (y > 0 && x < m_width-1)         { val += (*this)[y-1][x+1]  *16;middle += 16;}
            if (y > 0 && x < m_width-2)         { val += (*this)[y-1][x+2]  *4; middle += 4;}

            if (y > 1)                          { val += (*this)[y-2][x]    *7; middle += 7;}
            if (y > 1 && x > 0)                 { val += (*this)[y-2][x-1]  *4; middle += 4;}
            if (y > 1 && x > 1)                 { val += (*this)[y-2][x-2]  *1; middle += 1;}
            if (y > 1 && x < m_width-1)         { val += (*this)[y-2][x+1]  *4; middle += 4;}
            if (y > 1 && x < m_width-2)         { val += (*this)[y-2][x+2]  *1; middle += 1;}

            if (y < m_height-1)                 { val += (*this)[y+1][x]    *26;middle += 26;}
            if (y < m_height-1 && x > 0)        { val += (*this)[y+1][x-1]  *16;middle += 16;}
            if (y < m_height-1 && x > 1)        { val += (*this)[y+1][x-2]  *4; middle += 4;}
            if (y < m_height-1 && x < m_width-1){ val += (*this)[y+1][x+1]  *16;middle += 16;}
            if (y < m_height-1 && x < m_width-2){ val += (*this)[y+1][x+2]  *4; middle += 4;}

            if (y < m_height-2)                 { val += (*this)[y+2][x]    *7; middle += 7;}
            if (y < m_height-2 && x > 0)        { val += (*this)[y+2][x-1]  *4; middle += 4;}
            if (y < m_height-2 && x > 1)        { val += (*this)[y+2][x-2]  *1; middle += 1;}
            if (y < m_height-2 && x < m_width-1){ val += (*this)[y+2][x+1]  *4; middle += 4;}
            if (y < m_height-2 && x < m_width-2){ val += (*this)[y+2][x+2]  *1; middle += 1;}

            (*result)[y][x] = (int)val/middle; //Calculating average value for given pixel
        }
    }

    return [ImageWrapper imageWithCPPImage:result]; //Return the image
}

Hi Im new to image processing.

I would like to know if it is possible to modify the existing code to calculate 
the edge gradient and direction and then shade the edges appropriately.

The attached file show an example of what Im looking to achieve. 

NSPoint and NSRect are not available on the device.

I've temporarily typedeffed them to CGPoint and CGRect.

This should work on the device - we'll need to work out a way of switching
in the definitions for Simulator builds and unit test builds.

What steps will reproduce the problem?
1.
2.
3.

What is the expected output? What do you see instead?


Please use labels and text to provide additional information.

sorry to be dense, but how do I do that?

It would be good it the C++ code could all be moved behind the objective-C
wrapper. This would remove the need for having Objective-C++ in all the
project code.

The current way that bytes are converted back into an image could be improved.