See http://en.wikipedia.org/wiki/Polar_coordinate_system
There is probably a version in your language...

Note: I have no idea of what you mean by "change an image" in this context.

Also check out this blog post:

http://amnonp5.wordpress.com/2011/08/21/striate-cortex-advanced-pixel-manipulation/

that good work, you get nice effects!  amnon








But how could reprentar these coordinates as follows:??

sorry,  amnon ., and polar coordinates to Cartesian coordinates.??

Hi, I have created a code to pass, Cartesian coordinates to polar coordinates, the results below. I would like if you can take a look at the code, something wrong or if this could optimize better. if you see this half wavy image.

PS: I based code of  amnon

1. PImage input = loadImage("C:/Users/Naruto/Desktop/1.jpg");
     PImage output = createImage(input.width, input.height, RGB);
   output.loadPixels();
   for(float r=0;r<(input.height/2);r++){
   
     for(float q=0;q<TWO_PI;q=0.01+q){
    
     int polarX = int(r * cos(q)) + input.width/2;
     int polarY = int(r * sin(q)) + input.height/2;
    
       float x = map(q, TWO_PI, 0, 0, input.width);
   
         int outputIndex = int( x )+ int(r) * output.width;
         int inputIndex = polarX + polarY *input.width;
   
         output.pixels[outputIndex] = input.pixels[inputIndex];
     }
   
   }
   
   output.updatePixels();
   size( output.width, output.height);
   image(output,0,0);
   save("Out-imageCartesian.jpg");
   
   
   

Seems to work quite well.

Perhaps the following can be used instead of map...

1.     float x = (1-q/TWO_PI)*input.width;

Also, your code seems close to the first example, but the second (output-based) and third (LUTs) were actually more efficient. However I always like to say, whatever works. For only a few or static images this is fine, for series or realtime processing, you can move calculations to look-up tables (see the blog post's third example) or use GLSL shaders (which by now have been integrated quite well into Processing 2.0).

amnon , you can look at this page, the result is not the same as mine. my image is more compressed.
http://ajmas.dyndns.org/?loc=articles/programming/java/polar2cartesian


ps: I copied this "    float x = (1-q/TWO_PI)*input.width; " but not much difference.







my result:

Yeah, the suggested code wasn't different. It was identical, except (I think) less cpu-intensive than the map() method. The code you link to is concise and already in java. It seems possible to translate it into Processing fairly easily, if you are going after that result.

I think I've found the key.
Look at the results, I THINK THIS PERFECT
.


  Reviewer?


Is best to choose an image with width and height ODD To find your best ..

1. PImage input = loadImage("C:/Users/Naruto/Desktop/12.jpg");
   int anchosalida= int(input.height*PI);
   int altosalida= int(input.height);
     PImage output = createImage(anchosalida, altosalida, RGB);
   
    
    
   
   
   output.loadPixels();
   for(float r=0;r<input.height/2;r=r+0.01){
   
     for(float q=0;q<TWO_PI;q=0.01+q){
    
     int polarX = int(r * cos(q)) + input.width/2;
     int polarY = int(r * sin(q)) + input.height/2;
    
      float xa = map(q, 0, TWO_PI, 20, input.width*PI);
         //  float x = (1-q/TWO_PI)*input.width;
         int outputIndex = int(xa)+ int(r) * output.width;
         int inputIndex = polarX + polarY *input.width;
   
         output.pixels[outputIndex] = input.pixels[inputIndex];
     }
   
   }
   
   output.updatePixels();
   println(output.width);
   println(output.height);
   size( output.width, output.height,P2D);
   
   image(output,0,0);
   save("Out-imageCartesian.jpg");