load() is slow. Only load what you have to.
result=zeros(25,16000);
for K = 1 : length(vector1) % length is 25
temp1 = vector1(K);
all_t1{K} = load(matfiles1(temp1).name) %load image1 from matfile1
end
for K = 1 : length(vector2) % length is 16000
temp2 = vector2(K);
all_t2{K} = load(matfiles2(temp2).name) % load image2 from matfile2
end
for K1 = 1 : length(vector1)
t1 = all_t1{K1};
for K2 = 1 : length(vector2)
t2 = all_t2{K2};
result(i,j) = t1 .* t2;
end
end
Note: the fact that you store the element-by-element multiplication into a single numeric entry implies that the files you load might all consist of single scalar values. If that is the case, then the storage use can be improved and the multiplication can be sped up. Store the values into row vectors and then,
result = bsxfun(@times, all_t1(:), all_t2 )
