Info
此问题已关闭。 请重新打开它进行编辑或回答。
vectorization in context of if-condition and comparison for image array
1 次查看(过去 30 天)
显示 更早的评论
I want to implement some image compositing methods in Matlab like for example ModulusAdd defined as
if (Sca + Dca) <= 1
Dca = Sca + Dca;
else
Dca = (Sca + Dca) - 1;
end
Here Sca, Dca, Sa, Da are (h,w,1) arrays (gray scale images respective alpha channels) all as double (see also http://www.w3.org/TR/2014/CR-compositing-1-20140220/ for image compositing details).
Comparisons with the if-condition are defined for scalar values so using the above code in Matlab is not working directly (there are also other compositing methods that relay on other if-conditions like "if Sca == 0 && Dca == Da ..." that can not directly be ported so answering this question might also be helpful for those cases). Naively one has to deal with the pixel level and its two nested for-loops:
for i=1:h
for j=1:w
if (Sca(i,j) + Dca(i,j)) <= 1
Dca(i,j) = Sca(i,j) + Dca(i,j);
else
Dca(i,j) = (Sca(i,j) + Dca(i,j)) - 1;
end
end
end
How can I redesign this to make it work faster with an (h,w,1) (or (h,w,3)) array? Perhaps a redesign beyond arrayfun is possible (being not an expert in this):
Dca = arrayfun(@ModulusAdd_scalar, Sca, Dca);
function d = ModulusAdd_scalar(s, d)
if (s + d) <= 1
d = s + d;
else
d = (s + d) - 1;
end
end
0 个评论
回答(0 个)
此问题已关闭。
另请参阅
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!