Convert complex double to double type

Question

ana ain 2016-5-24

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/285594-convert-complex-double-to-double-type

回答： fengwei wang 2018-10-9

I am doing a feature extraction using glcm method. The glcm output are in 'struct' type, while i need the output in double type.

So, I've tried to convert it using several code showed below.

srcFile = dir('D:\1. Thesis FINISH!!!\*.png');
fetrain1 = [];
for b = 1:length(srcFile)
  file_name = strcat('D:\1. Thesis FINISH!!!\',srcFile(b).name);
  B = imread(file_name);
  [gl] = glcm (B);
  [gl] = struct2cell (gl);
  [gl] = cell2mat (gl);
%   [fiturtrain] = structfun(@double, gl, 'uniformoutput', 0);
%   [fiturtrain] = double (gl);
  [fiturtrain] = cell2mat (gl);
  fetrain1 = [fetrain1 fiturtrain];
%   vectorname = strcat(file_name,'_array.mat');
end
save ('fetrain1.mat','fetrain1');

But, the code give me 'complex double' type. I don't know why?

I need to convert fetrain1 from 'complex double' type into 'double' type, so I can use the variable for the next step. Or another way to covert 'struct' exactly to 'double' type.

Any suggestion would be very appreciated. Thank you.

4 个评论
显示 2更早的评论隐藏 2更早的评论

Megha 2017-7-10

It would be good enough to go with real(data) which could be real(fetrain1) in your case.

Hope this helps!

Walter Roberson 2017-9-18

Lin Li comments to Megha:

It helps to me. Thanks.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Andrew Newell 2017-4-19

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/285594-convert-complex-double-to-double-type#answer_263519

在 MATLAB Online 中打开

Judging by the discussion in the Stack Exchange, it seems likely that you can get what you want by replacing most of the code after B = imread(file_name) by the following:

imG = rgb2gray(B);
% create gray-level co-occurrence matrix from image
gimg = graycomatrix(imG);
% extract properties of gray-level co-occurrence matrix
stats = graycoprops(gimg, 'all');
features = [stats.Energy stats.Homogeneity stats.Correlation stats.Contrast];