Related
For a given vector $(x_1,x_2,\ldots, x_n)$ I am trying to compute
I wrote the following code
for l = 1:n
for k = 1:n
error = error + norm(x(i)-x(j))
end
end
This code is not fast, especially when $n$ is large. I am aware that I am double counting actually... But how may I avoid it? How can I speed up my code?
Thank you!
You can do it with bsxfun, which is fast:
d = (abs(bsxfun(#minus, x, x.')));
result = sum(d(:));
Or alternatively use pdist with 'cityblock' distance (which for one-dimensional observations reduces to absolute difference). This computes each distance once, so you need to multiply the sum by 2:
result = 2*sum(pdist(x(:),'cityblock'));
How about a simple speed up?
for a=1:n
for b=a+1:n
error = error + 2*norm(x(a)-x(b))
end
end
For a scalar, norm just gives abs.
So,
error = sum(abs( bsxfun(#minus, error,error') ))
will do the same thing.
also check out pdist which will do this for vectors, using vector norms, in an even faster way.
I've written code to smooth an image using a 3x3 averaging filter, however the output is strange, it is almost all black. Here's my code.
function [filtered_img] = average_filter(noisy_img)
[m,n] = size(noisy_img);
filtered_img = zeros(m,n);
for i = 1:m-2
for j = 1:n-2
sum = 0;
for k = i:i+2
for l = j:j+2
sum = sum+noisy_img(k,l);
end
end
filtered_img(i+1,j+1) = sum/9.0;
end
end
end
I call the function as follows:
img=imread('img.bmp');
filtered = average_filter(img);
imshow(uint8(filtered));
I can't see anything wrong in the code logic so far, I'd appreciate it if someone can spot the problem.
Assuming you're working with grayscal images, you should replace the inner two for loops with :
filtered_img(i+1,j+1) = mean2(noisy_img(i:i+2,j:j+2));
Does it change anything?
EDIT: don't forget to reconvert it to uint8!!
filtered_img = uint8(filtered_img);
Edit 2: the reason why it's not working in your code is because sum is saturating at 255, the upper limit of uint8. mean seems to prevent that from happening
another option:
f = #(x) mean(x(:));
filtered_img = nlfilter(noisy_img,[3 3],f);
img = imread('img.bmp');
filtered = imfilter(double(img), ones(3) / 9, 'replicate');
imshow(uint8(filtered));
Implement neighborhood operation of sum of product operation between an image and a filter of size 3x3, the filter should be averaging filter.
Then use the same function/code to compute Laplacian(2nd order derivative, prewitt and sobel operation(first order derivatives).
Use a simple 10*10 matrix to perform these operations
need matlab code
Tangentially to the question:
Especially for 5x5 or larger window you can consider averaging first in one direction and then in the other and you save some operations. So, point at 3 would be (P1+P2+P3+P4+P5). Point at 4 would be (P2+P3+P4+P5+P6). Divided by 5 in the end. So, point at 4 could be calculated as P3new + P6 - P2. Etc for point 5 and so on. Repeat the same procedure in other direction.
Make sure to divide first, then sum.
I would need to time this, but I believe it could work a bit faster for larger windows. It is sequential per line which might not seem the best, but you have many lines where you can work in parallel, so it shouldn't be a problem.
This first divide, then sum also prevents saturation if you have integers, so you might use the approach even in 3x3 case, as it is less wrong (though slower) to divide twice by 3 than once by 9. But note that you will always underestimate final value with that, so you might as well add a bit of bias (say all values +1 between the steps).
img=imread('camraman.tif');
nsy-img=imnoise(img,'salt&pepper',0.2);
imshow('nsy-img');
h=ones(3,3)/9;
avg=conv2(img,h,'same');
imshow(Unit8(avg));
Several posts exist about efficiently calculating pairwise distances in MATLAB. These posts tend to concern quickly calculating euclidean distance between large numbers of points.
I need to create a function which quickly calculates the pairwise differences between smaller numbers of points (typically less than 1000 pairs). Within the grander scheme of the program i am writing, this function will be executed many thousands of times, so even small gains in efficiency are important. The function needs to be flexible in two ways:
On any given call, the distance metric can be euclidean OR city-block.
The dimensions of the data are weighted.
As far as i can tell, no solution to this particular problem has been posted. The statstics toolbox offers pdist and pdist2, which accept many different distance functions, but not weighting. I have seen extensions of these functions that allow for weighting, but these extensions do not allow users to select different distance functions.
Ideally, i would like to avoid using functions from the statistics toolbox (i am not certain the user of the function will have access to those toolboxes).
I have written two functions to accomplish this task. The first uses tricky calls to repmat and permute, and the second simply uses for-loops.
function [D] = pairdist1(A, B, wts, distancemetric)
% get some information about the data
numA = size(A,1);
numB = size(B,1);
if strcmp(distancemetric,'cityblock')
r=1;
elseif strcmp(distancemetric,'euclidean')
r=2;
else error('Function only accepts "cityblock" and "euclidean" distance')
end
% format weights for multiplication
wts = repmat(wts,[numA,1,numB]);
% get featural differences between A and B pairs
A = repmat(A,[1 1 numB]);
B = repmat(permute(B,[3,2,1]),[numA,1,1]);
differences = abs(A-B).^r;
% weigh difference values before combining them
differences = differences.*wts;
differences = differences.^(1/r);
% combine features to get distance
D = permute(sum(differences,2),[1,3,2]);
end
AND:
function [D] = pairdist2(A, B, wts, distancemetric)
% get some information about the data
numA = size(A,1);
numB = size(B,1);
if strcmp(distancemetric,'cityblock')
r=1;
elseif strcmp(distancemetric,'euclidean')
r=2;
else error('Function only accepts "cityblock" and "euclidean" distance')
end
% use for-loops to generate differences
D = zeros(numA,numB);
for i=1:numA
for j=1:numB
differences = abs(A(i,:) - B(j,:)).^(1/r);
differences = differences.*wts;
differences = differences.^(1/r);
D(i,j) = sum(differences,2);
end
end
end
Here are the performance tests:
A = rand(10,3);
B = rand(80,3);
wts = [0.1 0.5 0.4];
distancemetric = 'cityblock';
tic
D1 = pairdist1(A,B,wts,distancemetric);
toc
tic
D2 = pairdist2(A,B,wts,distancemetric);
toc
Elapsed time is 0.000238 seconds.
Elapsed time is 0.005350 seconds.
Its clear that the repmat-and-permute version works much more quickly than the double-for-loop version, at least for smaller datasets. But i also know that calls to repmat often slow things down, however. So I am wondering if anyone in the SO community has any advice to offer to improve the efficiency of either function!
EDIT
#Luis Mendo offered a nice cleanup of the repmat-and-permute function using bsxfun. I compared his function with my original on datasets of varying size:
As the data become larger, the bsxfun version becomes the clear winner!
EDIT #2
I have finished writing the function and it is available on github [link]. I ended up finding a pretty good vectorized method for computing euclidean distance [link], so i use that method in the euclidean case, and i took #Divakar's advice for city-block. It is still not as fast as pdist2, but its must faster than either of the approaches i laid out earlier in this post, and easily accepts weightings.
You can replace repmat by bsxfun. Doing so avoids explicit repetition, therefore it's more memory-efficient, and probably faster:
function D = pairdist1(A, B, wts, distancemetric)
if strcmp(distancemetric,'cityblock')
r=1;
elseif strcmp(distancemetric,'euclidean')
r=2;
else
error('Function only accepts "cityblock" and "euclidean" distance')
end
differences = abs(bsxfun(#minus, A, permute(B, [3 2 1]))).^r;
differences = bsxfun(#times, differences, wts).^(1/r);
D = permute(sum(differences,2),[1,3,2]);
end
For r = 1 ("cityblock" case), you can use bsxfun to get elementwise subtractions and then use matrix-multiplication, which must speed up things. The implementation would look something like this -
%// Calculate absolute elementiwse subtractions
absm = abs(bsxfun(#minus,permute(A,[1 3 2]),permute(B,[3 1 2])));
%// Perform matrix multiplications with the given weights and reshape
D = reshape(reshape(absm,[],size(A,2))*wts(:),size(A,1),[]);
Suppose that f(x,y) is a bivariate function as follows:
function [ f ] = f(x,y)
UN=(g)1.6*(1-acos(g)/pi)-0.8;
f= 1+UN(cos(0.5*pi*x+y));
end
How to improve execution time for function F(N) with the following code:
function [VAL] = F(N)
x=0:4/N:4;
y=0:2*pi/1000:2*pi;
VAL=zeros(N+1,3);
for i = 1:N+1
val = zeros(1,N+1);
for j = 1:N+1
val(j) = trapz(y,f(0,y).*f(x(i),y).*f(x(j),y))/2/pi;
end
val = fftshift(fft(val))/N;
l = (length(val)+1)/2;
VAL(i,:)= val(l-1:l+1);
end
VAL = fftshift(fft(VAL,[],1),1)/N;
L = (size(VAL,1)+1)/2;
VAL = VAL(L-1:L+1,:);
end
Note that N=2^p where p>10, so please consider the memory limitations while optimizing the code using ndgrid, arrayfun, etc.
FYI: The code intends to find the central 3-by-3 submatrix of the fftn of
fun=#(a,b) trapz(y,f(0,y).*f(a,y).*f(b,y))/2/pi;
where a,b are in [0,4]. The key idea is that we can save memory using the code above specially when N is very large. But the execution time is still an issue because of nested loops. See the figure below for N=2^2:
This is not a full answer, but some possibly helpful hints:
0) The trivial: Are you sure you need numerics? Can't you do the computation analytically?
1) Do not use function handles:
function [ f ] = f(x,y)
f= 1+1.6*(1-acos(cos(0.5*pi*x+y))/pi)-0.8
end
2) Simplify analytically: acos(cos(x)) is the same as abs(mod(x + pi, 2 * pi) - pi), which should compute slightly faster. Or, instead of sampling and then numerically integrating, first integrate analytically and sample the result.
3) The FFT is a very efficient algorithm to compute the full DFT, but you don't need the full DFT. Since you only want the central 3 x 3 coefficients, it might be more efficient to directly apply the DFT definition and evaluate the formula only for those coefficients that you want. That should be both fast and memory-efficient.
4) If you repeatedly do this computation, it might be helpful to precompute DFT coefficients. Here, dftmtx from the Signal Processing toolbox can assist.
5) To get rid of the loops, think about the problem not in the form of computation instructions, but a single matrix operation. If you consider your input N x N matrix as a vector with N² elements, and your output 3 x 3 matrix as a 9-element vector, then the whole operation you apply (numerical integration via trapz and DFT via fft) appears to be a simple linear transform, which it should be possible to express as an N² x 9 matrix.
Is there any efficient techniques to do the following summation ?
Given a finite set A containing n integers A={X1,X2,…,Xn}, where Xi is an integer. Now there are n subsets of A, denoted by A1, A2, ... , An. We want to calculate the summation for each subset. Are there some efficient techniques ?
(Note that n is typically larger than the average size of all the subsets of A.)
For example, if A={1,2,3,4,5,6,7,9}, A1={1,3,4,5} , A2={2,3,4} , A3= ... . A naive way of computing the summation for A1 and A2 needs 5 Flops for additions:
Sum(A1)=1+3+4+5=13
Sum(A2)=2+3+4=9
...
Now, if computing 3+4 first, and then recording its result 7, we only need 3 Flops for addtions:
Sum(A1)=1+7+5=13
Sum(A2)=2+7=9
...
What about the generalized case ? Is there any efficient methods to speed up the calculation? Thanks!
For some choices of subsets there are ways to speed up the computation, if you don't mind doing some (potentially expensive) precomputation, but not for all. For instance, suppose your subsets are {1,2}, {2,3}, {3,4}, {4,5}, ..., {n-1,n}, {n,1}; then the naive approach uses one arithmetic operation per subset, and you obviously can't do better than that. On the other hand, if your subsets are {1}, {1,2}, {1,2,3}, {1,2,3,4}, ..., {1,2,...,n} then you can get by with n-1 arithmetic ops, whereas the naive approach is much worse.
Here's one way to do the precomputation. It will not always find optimal results. For each pair of subsets, define the transition cost to be min(size of symmetric difference, size of Y - 1). (The symmetric difference of X and Y is the set of things that are in X or Y but not both.) So the transition cost is the number of arithmetic operations you need to do to compute the sum of Y's elements, given the sum of X's. Add the empty set to your list of subsets, and compute a minimum-cost directed spanning tree using Edmonds' algorithm (http://en.wikipedia.org/wiki/Edmonds%27_algorithm) or one of the faster but more complicated variations on that theme. Now make sure that when your spanning tree has an edge X -> Y you compute X before Y. (This is a "topological sort" and can be done efficiently.)
This will give distinctly suboptimal results when, e.g., you have {1,2}, {3,4}, {1,2,3,4}, {5,6}, {7,8}, {5,6,7,8}. After deciding your order of operations using the procedure above you could then do an optimization pass where you find cheaper ways to evaluate each set's sum given the sums already computed, and this will probably give fairly decent results in practice.
I suspect, but have made no attempt to prove, that finding an optimal procedure for a given set of subsets is NP-hard or worse. (It is certainly computable; the set of possible computations you might do is finite. But, on the face of it, it may be awfully expensive; potentially you might be keeping track of about 2^n partial sums, be adding any one of them to any other at each step, and have up to about n^2 steps, for a super-naive cost of (2^2n)^(n^2) = 2^(2n^3) operations to try every possibility.)
Assuming that 'addition' isn't simply an ADD operation but instead some very intensive function involving two integer operands, then an obvious approach would be to cache the results.
You could achieve that via a suitable data structure, for example a key-value dictionary containing keys formed by the two operands and the answers as the value.
But as you specified C in the question, then the simplest approach would be an n by n array of integers, where the solution to x + y is stored at array[x][y].
You can then repeatedly iterate over the subsets, and for each pair of operands you check the appropriate position in the array. If no value is present then it must be calculated and placed in the array. The value then replaces the two operands in the subset and you iterate.
If the operation is commutative then the operands should be sorted prior to looking up the array (i.e. so that the first index is always the smallest of the two operands) as this will maximise "cache" hits.
A common optimization technique is to pre-compute intermediate results. In your case, you might pre-compute all sums with 2 summands from A and store them in a lookup table. This will result in |A|*|A+1|/2 table entries, where |A| is the cardinality of A.
In order to compute the element sum of Ai, you:
look up the sum of the first two elements of Ai and save them in tmp
while there is an element x left in Ai:
look up the sum of tmp and x
In order to compute the element sum of A1 = {1,3,4,5} from your example, you do the following:
lookup(1,3) = 4
lookup(4,4) = 8
lookup(8,5) = 13
Note that computing the sum of any given Ai doesn't require summation, since all the work has already been conducted while pre-computing the lookup table.
If you store the lookup table in a hash table, then lookup() is in O(1).
Possible optimizations to this approach:
construct the lookup table while computing the summation results; hence, you only compute those summations that you actually need. Your lookup table is now a cache.
if your addition operation is commutative, you can save half of your cache size by storing only those summations where the smaller summand comes first. Then modify lookup() such that lookup(a,b) = lookup(b,a) if a > b.
If assuming summation is time consuming action you can find LCS of every pair of subsets (by assuming they are sorted as mentioned in comments, or if they are not sorted sort them), after that calculate sum of LCS of maximum length (over all LCS in pairs), then replace it's value in related arrays with related numbers, update their LCS and continue this way till there is no LCS with more than one number. Sure this is not optimum, but it's better than naive algorithm (smaller number of summation). However you can do backtracking to find best solution.
e.g For your sample input:
A1={1,3,4,5} , A2={2,3,4}
LCS (A_1,A_2) = {3,4} ==>7 ==>replace it:
A1={1,5,7}, A2={2,7} ==> LCS = {7}, maximum LCS length is `1`, so calculate sums.
Still you can improve it by calculation sum of two random numbers, then again taking LCS, ...
NO. There is no efficient techique.
Because it is NP complete problem. and there are no efficient solutions for such problem
why is it NP-complete?
We could use algorithm for this problem to solve set cover problem, just by putting extra set in set, conatining all elements.
Example:
We have sets of elements
A1={1,2}, A2={2,3}, A3 = {3,4}
We want to solve set cover problem.
we add to this set, set of numbers containing all elements
A4 = {1,2,3,4}
We use algorhitm that John Smith is aking for and we check solution A4 is represented whit.
We solved NP-Complete problem.