How to combine normxcorr2 similarity results with single value, like mutual information?

1 回表示 (過去 30 日間)
Mohammad Al Nagdawi
Mohammad Al Nagdawi 2017 年 11 月 1 日
編集済み: Mohammad Al Nagdawi 2017 年 11 月 2 日
Hi everyone
I want a single value normalized cross-correlation similarity between two images while transforming. This value I want to compare with other similarity measures, such as mutual information.
for i = 50:step:end_rang
% horizantal transformation
[overlap_fix,overlap_float] = overlapArea(fix_img,float_img,i);
MI(2,ind) = size(overlap_fix,2);
MI(3,ind) = MutualInfo(overlap_fix,overlap_float,all_float_px,overlap_limit);
JD(ind) = jeffrey_divergence_Xu_paper(overlap_fix,overlap_float,all_float_px,overlap_limit);
if std(double(overlap_fix(:))) ~= 0
cc(ind) = sum(sum(normxcorr2(overlap_fix,overlap_float)))/numel(overlap_fix);
else
cc(ind) = 0;
disp(['skip ',num2str(i)]);
end
ind = ind +1;
end
The method I use to convert 'normxcorr2' output matrix to a single value, is it correct? where the identical images have the highest value

回答 (0 件)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by