matlab - How to prepare GLCM features for a neural network? -
i have used glcm in matlab extract texture features image. did using:
offsets0=[0 1; -1 1; -1;0, -1 1]; glcms=graycomatrix(img, 'offset', offsets0); stats=graycoprops(glcms, 'all'); figure(2), plot([stats.correlation]) % repeated contrast, homogeneity , energy.
this produces values:
homgeneity= 0.989 0.1468 0.1144 0.1342
energy= 0.3366 0.3281 0.3340 0.3296
contrast= 0.9542 0.9353 0.9474 0.9384
correlation= 0.9798 0.9699 0.9766 0.9707
i want use nnprtool gui classify these features unsure how prepare these features input neural network.
you might know already, can call gui using nnstart
there option @ end of gui save script created during gui. in script, see matlab can pre-processing functions such removing constant rows (removeconstantrows) , mapping row minimum , maximum values [-1 1] (mapminmax). hope helpful.
Comments
Post a Comment