matlab 函數名稱： graycoprops()
功 能： 計算灰度共生矩陣(GLCM)的各個特徵值+
句法：
stats = graycoprops(glcm, properties)
描述：
stats = graycoprops(glcm, properties) 從灰度共生矩陣(glcm)中，計算properties指定的特徵的統計值。glcm是一個m x n x p 的矩陣，它由有效的灰度共生矩陣(m x n)組成。如果glcm是一個GLCMs的數組，stats是由每個glcm的統計值組成的矩陣。
graycoprops() 歸一化灰度共生矩陣，以便它的元素的和等於1。歸一化的GLCM中的每一個元素(r,c)是圖像中像素值分別爲r和c的像素對出現的聯合概率。graycoprops使用歸一化的GLCM來計算特徵值。
properties可以是由逗號隔開的字符串，字符元胞數組，'all’字符串，或者空格隔開的字符串。
property的取值：
'Contrast' 即對比度，返回整幅圖像中某個像素與它的鄰居之間的對比度。常量組成的圖像的Contrast是0。計算公式爲：
'Correlation' 即互相關，返回整幅圖像中某個像素與它的鄰居之間的互相關度。取值範圍是 [-1 , 1]。常量組成的圖像的互相關度Correlation是NaN。相關度1和-1分別對應完全正相關和完全負相關。計算公式爲：
'Energy' 即能量，返回GLCM中所有元素的平方和。取值範圍是 [0 , 1]。常量組成的圖像的能量Energy是1。計算公式爲：
'Homogeneity' 即齊次性(同質性)，返回的值反映了GLCM中元素相對於GLCM對角線的分佈的緊密度。取值範圍是 [0 , 1]。對角的GLCM的Homogeneity齊次性是1。公式爲：
graycoprops()的返回變量states是一個結構體，它的域fields由properties指定。每一個field包含一個 1 x p 的數組，p是GLCM中灰度共生矩陣的數目。例如，GLCM是一個8 x 8 x 3 的矩陣，並且屬性爲‘Energy’，那麼，states是一個包含‘Energy’域的結構體，其中，Energy包含一個 1 x 3 的數組。
註釋：
Energy(能量)也叫做uniformity, uniformity of energy, 和 angular second moment。
Contrast(對比度)也叫做variance 和 inertia。
支持的類型：
glcm可以是邏輯型或數值型，並且它必須由非負的、有限的（和實數的）整數組成。 states是一個結構體。
舉例說明：
GLCM = [0 1 2 3;1 1 2 3;1 0 2 0;0 0 0 3];
stats = graycoprops(GLCM)
I = imread('circuit.tif');
GLCM2 = graycomatrix(I,'Offset',[2 0;0 2]);
stats = graycoprops(GLCM2,{'contrast','homogeneity'})
************************************************************************************************************************************************************************
************************************************************* graycoprops源程序代碼 *****************************************************************************
************************************************************************************************************************************************************************
function stats = graycoprops(varargin)
%GRAYCOPROPS Properties of gray-level co-occurrence matrix.
% STATS = GRAYCOPROPS(GLCM,PROPERTIES) normalizes the gray-level
% co-occurrence matrix (GLCM) so that the sum of its elements is one. Each
% element in the normalized GLCM, (r,c), is the joint probability occurrence
% of pixel pairs with a defined spatial relationship having gray level
% values r and c in the image. GRAYCOPROPS uses the normalized GLCM to
% calculate PROPERTIES.
%
% GLCM can be an m x n x p array of valid gray-level co-occurrence
% matrices. Each gray-level co-occurrence matrix is normalized so that its
% sum is one.
%
% PROPERTIES can be a comma-separated list of strings, a cell array
% containing strings, the string 'all', or a space separated string. They
% can be abbreviated, and case does not matter.
%
% Properties include:
%
% 'Contrast' the intensity contrast between a pixel and its neighbor
% over the whole image. Range = [0 (size(GLCM,1)-1)^2].
% Contrast is 0 for a constant image.
%
% 'Correlation' statistical measure of how correlated a pixel is to its
% neighbor over the whole image. Range = [-1 1].
% Correlation is 1 or -1 for a perfectly positively or
% negatively correlated image. Correlation is NaN for a
% constant image.
%
% 'Energy' summation of squared elements in the GLCM. Range = [0 1].
% Energy is 1 for a constant image.
%
% 'Homogeneity' closeness of the distribution of elements in the GLCM to
% the GLCM diagonal. Range = [0 1]. Homogeneity is 1 for
% a diagonal GLCM.
%
% If PROPERTIES is the string 'all', then all of the above properties are
% calculated. This is the default behavior. Please refer to the
% Documentation for more information on these properties.
%
% STATS is a structure with fields that are specified by PROPERTIES. Each
% field contains a 1 x p array, where p is the number of gray-level
% co-occurrence matrices in GLCM. For example, if GLCM is an 8 x 8 x 3 array
% and PROPERTIES is 'Energy', then STATS is a structure containing the
% field 'Energy'. This field contains a 1 x 3 array.
%
% Notes
% -----
% Energy is also known as uniformity, uniformity of energy, and angular second
% moment.
%
% Contrast is also known as variance and inertia.
%
% Class Support
% -------------
% GLCM can be logical or numeric, and it must contain real, non-negative, finite,
% integers. STATS is a structure.
%
% Examples
% --------
% GLCM = [0 1 2 3;1 1 2 3;1 0 2 0;0 0 0 3];
% stats = graycoprops(GLCM)
%
% I = imread('circuit.tif');
% GLCM2 = graycomatrix(I,'Offset',[2 0;0 2]);
% stats = graycoprops(GLCM2,{'contrast','homogeneity'})
%
% See also GRAYCOMATRIX.
% Copyright 2003-2009 The MathWorks, Inc.
allStats = {'Contrast','Correlation','Energy','Homogeneity'};
[glcm, requestedStats] = ParseInputs(allStats, varargin{:});
% Initialize output stats structure.
numStats = length(requestedStats);
numGLCM = size(glcm,3);
empties = repmat({zeros(1,numGLCM)},[numStats 1]);
stats = cell2struct(empties,requestedStats,1);
for p = 1 : numGLCM
if numGLCM ~= 1 %N-D indexing not allowed for sparse.
tGLCM = normalizeGLCM(glcm(:,:,p));
else
tGLCM = normalizeGLCM(glcm);
end
% Get row and column subscripts of GLCM. These subscripts correspond to the
% pixel values in the GLCM.
s = size(tGLCM);
[c,r] = meshgrid(1:s(1),1:s(2));
r = r(:);
c = c(:);
% Calculate fields of output stats structure.
for k = 1:numStats
name = requestedStats{k};
switch name
case 'Contrast'
stats.(name)(p) = calculateContrast(tGLCM,r,c);
case 'Correlation'
stats.(name)(p) = calculateCorrelation(tGLCM,r,c);
case 'Energy'
stats.(name)(p) = calculateEnergy(tGLCM);
case 'Homogeneity'
stats.(name)(p) = calculateHomogeneity(tGLCM,r,c);
end
end
end
%-----------------------------------------------------------------------------
function glcm = normalizeGLCM(glcm)
% Normalize glcm so that sum(glcm(:)) is one.
if any(glcm(:))
glcm = glcm ./ sum(glcm(:));
end
%-----------------------------------------------------------------------------
function C = calculateContrast(glcm,r,c)
% Reference: Haralick RM, Shapiro LG. Computer and Robot Vision: Vol. 1,
% Addison-Wesley, 1992, p. 460.
k = 2;
l = 1;
term1 = abs(r - c).^k;
term2 = glcm.^l;
term = term1 .* term2(:);
C = sum(term);
%-----------------------------------------------------------------------------
function Corr = calculateCorrelation(glcm,r,c)
% References:
% Haralick RM, Shapiro LG. Computer and Robot Vision: Vol. 1, Addison-Wesley,
% 1992, p. 460.
% Bevk M, Kononenko I. A Statistical Approach to Texture Description of Medical
% Images: A Preliminary Study., The Nineteenth International Conference of
% Machine Learning, Sydney, 2002.
% http://www.cse.unsw.edu.au/~icml2002/workshops/MLCV02/MLCV02-Bevk.pdf, p.3.
% Correlation is defined as the covariance(r,c) / S(r)*S(c) where S is the
% standard deviation.
% Calculate the mean and standard deviation of a pixel value in the row
% direction direction. e.g., for glcm = [0 0;1 0] mr is 2 and Sr is 0.
mr = meanIndex(r,glcm);
Sr = stdIndex(r,glcm,mr);
% mean and standard deviation of pixel value in the column direction, e.g.,
% for glcm = [0 0;1 0] mc is 1 and Sc is 0.
mc = meanIndex(c,glcm);
Sc = stdIndex(c,glcm,mc);
term1 = (r - mr) .* (c - mc) .* glcm(:);
term2 = sum(term1);
Corr = term2 / (Sr * Sc);
%-----------------------------------------------------------------------------
function S = stdIndex(index,glcm,m)
term1 = (index - m).^2 .* glcm(:);
S = sqrt(sum(term1));
%-----------------------------------------------------------------------------
function M = meanIndex(index,glcm)
M = index .* glcm(:);
M = sum(M);
%-----------------------------------------------------------------------------
function E = calculateEnergy(glcm)
% Reference: Haralick RM, Shapiro LG. Computer and Robot Vision: Vol. 1,
% Addison-Wesley, 1992, p. 460.
foo = glcm.^2;
E = sum(foo(:));
%-----------------------------------------------------------------------------
function H = calculateHomogeneity(glcm,r,c)
% Reference: Haralick RM, Shapiro LG. Computer and Robot Vision: Vol. 1,
% Addison-Wesley, 1992, p. 460.
term1 = (1 + abs(r - c));
term = glcm(:) ./ term1;
H = sum(term);
%-----------------------------------------------------------------------------
function [glcm,reqStats] = ParseInputs(allstats,varargin)
numstats = length(allstats);
iptchecknargin(1,numstats+1,nargin,mfilename);
reqStats = '';
glcm = varargin{1};
% The 'nonnan' and 'finite' attributes are not added to iptcheckinput because the
% 'integer' attribute takes care of these requirements.
iptcheckinput(glcm,{'logical','numeric'},{'real','nonnegative','integer'}, ...
mfilename,'GLCM',1);
if ndims(glcm) > 3
eid = sprintf('Images:%s:invalidSizeForGLCM',mfilename);
msg = 'GLCM must either be an m x n or an m x n x p array.';
error(eid,'%s',msg);
end
% Cast GLCM to double to avoid truncation by data type. Note that GLCM is not an
% image.
if ~isa(glcm,'double')
glcm = double(glcm);
end
list = varargin(2:end);
if isempty(list)
% GRAYCOPROPS(GLCM) or GRAYCOPROPS(GLCM,PROPERTIES) where PROPERTIES is empty.
reqStats = allstats;
else
if iscell(list{1}) || numel(list) == 1
% GRAYCOPROPS(GLCM,{...})
list = list{1};
end
if ischar(list)
%GRAYCOPROPS(GLCM,SPACE-SEPARATED STRING)
list = strread(list, '%s');
end
anyprop = allstats;
anyprop{end+1} = 'all';
for k = 1 : length(list)
match = iptcheckstrs(list{k}, anyprop, mfilename, 'PROPERTIES', k+1);
if strcmp(match,'all')
reqStats = allstats;
break;
end
reqStats{k} = match;
end
end
% Make sure that reqStats are in alphabetical order.
reqStats = sort(reqStats);
if isempty(reqStats)
eid = sprintf('Images:%s:internalError',mfilename);
msg = 'Internal error: requestedStats has not been assigned.';
error(eid,'%s',msg);
end