flood_fill algorithm in Matlab - works, but doesnt stop

Question

Im developing a flood_fill algorithm in MATLAB right now, and I have a little problem that cost me a lot of time:

function [ homoPoints ] = area2Points( matrix )
%AREA2POINTS makes an area of one's to only one one in the middle of the area
% Input    :     Matrix with areas of one's
% Output   :     result has one point in the middle of every former area   

myMatrix = matrix;
[row, col] = find(myMatrix);
curPoint = [ row(1),col(1) ];
area = [curPoint(1),curPoint(2)];
myMatrix(curPoint(1),curPoint(2)) = 0;
myMatrix = fill(curPoint(1),curPoint(2),myMatrix);

%Problem
    function[ matrix ] = fill(x,y,matrix)
        area
        %matrix(x,y) = 0;    %theoretisch unnötig
        %If the pixel under curPoint is a 1..
        if(matrix(x + 1 , y) == 1)
            area = vertcat(area,[x+1 , y]);
            matrix(x+1,y) = 0;
            fill(x+1,y,matrix);
        end
        %If the pixel left from curPoint is a 1..
        if(matrix(x, y - 1) == 1)
            area = vertcat(area,[x , y-1]);
            matrix(x,y-1) = 0;
            fill(x,y-1,matrix);
        end
        %If the pixel over curPoint is a 1..
        if(matrix(x - 1, y) == 1)
            area = vertcat(area,[x-1 , y]);
            matrix(x-1,y) = 0;
            fill(x-1,y,matrix);
        end
        %If the pixel right from curPoint is a 1..
        if(matrix(x , y + 1) == 1)
            area = vertcat(area,[x , y+1]);
            matrix(x,y+1) = 0;
            fill(x,y+1,matrix);
        end
        return
    end

So the thing is: The flood_fill runs correctly through all pixels, but when all pixels are set 0, it doesnt stop! e.g. in this matrix:

testMatrix = zeros(20);
testMatrix(5:10,5:10) = 1;

..it goes down in the fith column, up in the six, down in the seventh.. up in the tenth (7,10 -> 6,10 -> 5,10) and then (8,10 -> 9,10 -> 10,10 -> 10,9 ...). Where does this effect come from?

score 0 · Answer 1 · answered Oct 07 '14 at 07:32

0

I'm sorry. I concentrated on the wrong thing.. I just solved it by accessing myMatrix directly, instead of the parameter. That worked fine!

Best regards, R93!

answered Oct 07 '14 at 07:32

Andreas Limberg

51
8

2

Answering your own question if you figure out how to do it is encouraged on Stack Overflow, but it would be nice if you can add some additional information explaining the solution with some more details. – Stewie Griffin Oct 07 '14 at 07:46

flood_fill algorithm in Matlab - works, but doesnt stop

1 Answers1