FINDLOCALMAX - Local maxima of a matrix.

Contents

Description

Determine the local maxima of a given matrix image.

Syntax

   max_local = FINDLOCALMAX(I);
   [max_local,row,col] = FINDLOCALMAX(I, radius, method );

Inputs

I : a (N,M) matrix containing values to filter.

*radius : optional variable setting the radius of the neighborhood; default: |radius=1, ie maxima in neighbourhoods (3,3) are extracted.

method : optional string defining the technique used for extracting the maxima; it is either 'dil' (fast, not unique), 'filt' (fast) or 'screen' (slow); default: method='filt'.

Outputs

row : the row position of the local maxima.

col : the column position of the local maxima.

max_local : the (N,M) matrix containing values of val on unique local maximum.

Remark

FINDLOCALMAX and FINDLOCALEXTREMA(..,'ord','max','corr','forall') provide identical results, except for the borders of the image.

See also

Ressembles: FINDLOCALEXTREMA, MAPTRANSITION, ORDFILT2. Requires: FINDLOCALMAX_BASE.

Function implementation

function [max_local,row,col] = findlocalmax(I,varargin)

parsing and checking parameters

if ~isnumeric(I)
    error('findlocalmax:inputerror','a matrix is required in input');
end

p = createParser('FINDLOCALMAX');
% mandatory parameter
% optional parameters
p.addOptional('radius',1., @(x)x>=0); % just for testing
p.addOptional('find', 'filt', @(x)ischar(x) && ...
    any(strcmpi(x,{'filt','dil','screen'})));

% parse and validate all input arguments
p.parse(varargin{:});
p = getvarParser(p);

dealing with multispectral images [M,N,C] = size(I); if C>1 max_local = zeros(size(I)); row = cell(C); col = cell(C); for c=1:C [max_local(:,:,c), row{c}, col{c}] = findlocalmax(I(:,:,c), 'parse', p); end return; end

main computation

[max_local,row,col] = findlocalmax_base(I, p.radius, p.find);

display

if p.disp
    figure, imagesc(max_local), axis image off, title('local maxima')
end

end % end of findlocalmax