For loop

hi 2 all, Can anyone say how to reduce execution time of 'for loop' in matlab How to replace for loop?

[Andrei Bobrov]

your cod with example

[harjan]

Thx a lot Here is my coding Pls say how to Change this "for loop"

[x y]=size(ima); %ima is 1600x1600 image

for i=1:x

for j=1:y

x(i,j)=sign(ima(i,j));

u1(i,j)=255*x(i,j);

...

...

end

end

It takes some amount of time to execute ....How to replace this Tell your suggestions......

Answer 1

[Andrei Bobrov]

u1 = 255*sign(ima);

but you get the array size is 1600 x 1600, with values of -255, 0 and 255

ADD on comments

if i1 and j1 integers from 1 to 1600, we obtain an array of 1600 by 1600, consisting of ones, in it case use

i1 = linspace(0,1,6);
j1 = i1;

T=cos(2*pi*repmat(i1',1,numel(j1))).*cos(2*pi*repmat(j1,numel(i1),1)); % 1 variant
T = bsxfun(@times,cos(2*pi*i1'),cos(2*pi*j1)); % 2  variant
T = cos(2*pi*i1')*cos(2*pi*j1);% 3 variant

[harjan]

But how to change if this is in for loop

T(i,j)=(cos(2*pi*i)) * (cos(2*pi*j)); %for i,j varies from 1 to 1600

Thx in advance

Answer 2

[Walter Roberson]

The one sure way to reduce the execution time of a for loop is to eliminate that section of code.

Anything else depends on exactly what your for loop contains. There are some for loops that are now faster than vectorizing or using one of the library routines -- faster even than using a mex routine.

[harjan]

Thx a lot Here is my coding Pls say how to Change this "for loop"

[x y]=size(ima); %ima is 1600x1600 image

for i=1:x

for j=1:y

x(i,j)=sign(ima(i,j));

u1(i,j)=255*x(i,j);

...

...

end

end

It takes some amount of time to execute ....How to replace this Tell your suggestions......