neighbour pixel operation

sir i have a white and black(BW) (binary)image of size (i X j).i need image to be scanned from left to right, up and down and the scan area is 3X3.if one of 8-neighbour pixel is white (255).then center point should becomes white (255). sir please help in coding

sir i am able to divide image into blocks of size 3X3

BW=imagen;

BW2 = zeros(3 * ceil(size(BW)/3));

BW2(1:size(BW,1),1:size(BW,2)) = BW;

BW = BW2;

r = mat2cell(BW, 3 * ones(1,size(BW,1)/3), 3 * ones(1,size(BW,2)/3));

then i am trying with B(2,2) = max(r(2,2),255 * any(r(:)==255)); %for (2 2)block

but i am geting error

 ??? Function 'eq' is not defined for values of class 'cell'.

Error in ==> eq at 18 [varargout{1:nargout}] = builtin('eq', varargin{:});

sir please help in coading

Answer 1

[Oleg Komarov]

Why did you group in cells in thr first place?

Referencing the content of a cell should be done with curly brackets:

r{1,1}

[javeed miyandad]

sir i used B{2,2} = max(r{2,2},255 * any(r{:}==255));

but geting error

??? Error using ==> eq

Too many input arguments.

[Oleg Komarov]

Try to execute r{:}. It calls the content of every cell SEQUENTIALLY and not all together.

That's why I asked you why did you use cells.

You could keep the whole double matrix and reference the blocks (r{2,2} in this case) with logical indexing.

[javeed miyandad]

sorry sir i am not geting

plz can u give another code

[Oleg Komarov]

B{2,2} = max(r{2,2}, 255*cellfun(@(x) any(x == 255),x));