I'm trying to write an image compression script in MATLAB using multilayer 3D DWT(color image). along the way, I want to apply thresholding on coefficient matrices, both global and local thresholds.
I like to use the formula below to calculate my local threshold:
where sigma is variance and N is the number of elements.
Global thresholding works fine; but my problem is that the calculated local threshold is (most often!) greater than the maximum band coefficient, therefore no thresholding is applied.
Everything else works fine and I get a result too, but I suspect the local threshold is miscalculated. Also, the resulting image is larger than the original!
I'd appreciate any help on the correct way to calculate the local threshold, or if there's a pre-set MATLAB function.
here's an example output:
here's my code:
clear;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%% COMPRESSION %%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% read base image
% dwt 3/5-L on base images
% quantize coeffs (local/global)
% count zero value-ed coeffs
% calculate mse/psnr
% save and show result
% read images
base = imread('circ.jpg');
fam = 'haar'; % wavelet family
lvl = 3; % wavelet depth
% set to 1 to apply global thr
thr_type = 0;
% global threshold value
gthr = 180;
% convert base to grayscale
%base = rgb2gray(base);
% apply dwt on base image
dc = wavedec3(base, lvl, fam);
% extract coeffs
ll_base = dc.dec{1};
lh_base = dc.dec{2};
hl_base = dc.dec{3};
hh_base = dc.dec{4};
ll_var = var(ll_base, 0);
lh_var = var(lh_base, 0);
hl_var = var(hl_base, 0);
hh_var = var(hh_base, 0);
% count number of elements
ll_n = numel(ll_base);
lh_n = numel(lh_base);
hl_n = numel(hl_base);
hh_n = numel(hh_base);
% find local threshold
ll_t = ll_var * (sqrt(2 * log2(ll_n)));
lh_t = lh_var * (sqrt(2 * log2(lh_n)));
hl_t = hl_var * (sqrt(2 * log2(hl_n)));
hh_t = hh_var * (sqrt(2 * log2(hh_n)));
% global
if thr_type == 1
ll_t = gthr; lh_t = gthr; hl_t = gthr; hh_t = gthr;
end
% count zero values in bands
ll_size = size(ll_base);
lh_size = size(lh_base);
hl_size = size(hl_base);
hh_size = size(hh_base);
% count zero values in new band matrices
ll_zeros = sum(ll_base==0,'all');
lh_zeros = sum(lh_base==0,'all');
hl_zeros = sum(hl_base==0,'all');
hh_zeros = sum(hh_base==0,'all');
% initiate new matrices
ll_new = zeros(ll_size);
lh_new = zeros(lh_size);
hl_new = zeros(lh_size);
hh_new = zeros(lh_size);
% apply thresholding on bands
% if new value < thr => 0
% otherwise, keep the previous value
for id=1:ll_size(1)
for idx=1:ll_size(2)
if ll_base(id,idx) < ll_t
ll_new(id,idx) = 0;
else
ll_new(id,idx) = ll_base(id,idx);
end
end
end
for id=1:lh_size(1)
for idx=1:lh_size(2)
if lh_base(id,idx) < lh_t
lh_new(id,idx) = 0;
else
lh_new(id,idx) = lh_base(id,idx);
end
end
end
for id=1:hl_size(1)
for idx=1:hl_size(2)
if hl_base(id,idx) < hl_t
hl_new(id,idx) = 0;
else
hl_new(id,idx) = hl_base(id,idx);
end
end
end
for id=1:hh_size(1)
for idx=1:hh_size(2)
if hh_base(id,idx) < hh_t
hh_new(id,idx) = 0;
else
hh_new(id,idx) = hh_base(id,idx);
end
end
end
% count zeros of the new matrices
ll_new_size = size(ll_new);
lh_new_size = size(lh_new);
hl_new_size = size(hl_new);
hh_new_size = size(hh_new);
% count number of zeros among new values
ll_new_zeros = sum(ll_new==0,'all');
lh_new_zeros = sum(lh_new==0,'all');
hl_new_zeros = sum(hl_new==0,'all');
hh_new_zeros = sum(hh_new==0,'all');
% set new band matrices
dc.dec{1} = ll_new;
dc.dec{2} = lh_new;
dc.dec{3} = hl_new;
dc.dec{4} = hh_new;
% count how many coeff. were thresholded
ll_zeros_diff = ll_new_zeros - ll_zeros;
lh_zeros_diff = lh_zeros - lh_new_zeros;
hl_zeros_diff = hl_zeros - hl_new_zeros;
hh_zeros_diff = hh_zeros - hh_new_zeros;
% show coeff. matrices vs. thresholded version
figure
colormap(gray);
subplot(2,4,1); imagesc(ll_base); title('LL');
subplot(2,4,2); imagesc(lh_base); title('LH');
subplot(2,4,3); imagesc(hl_base); title('HL');
subplot(2,4,4); imagesc(hh_base); title('HH');
subplot(2,4,5); imagesc(ll_new); title({'LL thr';ll_zeros_diff});
subplot(2,4,6); imagesc(lh_new); title({'LH thr';lh_zeros_diff});
subplot(2,4,7); imagesc(hl_new); title({'HL thr';hl_zeros_diff});
subplot(2,4,8); imagesc(hh_new); title({'HH thr';hh_zeros_diff});
% idwt to reconstruct compressed image
cmp = waverec3(dc);
cmp = uint8(cmp);
% calculate mse/psnr
D = abs(cmp - base) .^2;
mse = sum(D(:))/numel(base);
psnr = 10*log10(255*255/mse);
% show images and mse/psnr
figure
subplot(1,2,1);
imshow(base); title("Original"); axis square;
subplot(1,2,2);
imshow(cmp); colormap(gray); axis square;
msg = strcat("MSE: ", num2str(mse), " | PSNR: ", num2str(psnr));
title({"Compressed";msg});
% save image locally
imwrite(cmp, 'compressed.png');
I solved the question.
the sigma in the local threshold formula is not variance, it's the standard deviation. I applied these steps:
used stdfilt() std2() to find standard deviation of my coeff. matrices (thanks to #Rotem for pointing this out)
used numel() to count the number of elements in coeff. matrices
this is a summary of the process. it's the same for other bands (LH, HL, HH))
[c, s] = wavedec2(image, wname, level); %apply dwt
ll = appcoeff2(c, s, wname); %find LL
ll_std = std2(ll); %find standard deviation
ll_n = numel(ll); %find number of coeffs in LL
ll_t = ll_std * (sqrt(2 * log2(ll_n))); %local the formula
ll_new = ll .* double(ll > ll_t); %thresholding
replace the LL values in c in a for loop
reconstruct by applying IDWT using waverec2
this is a sample output:
Assume that I have two discrete random variables X and Y.
X = {1,3,3,5,7,7,7,9,9,9,9,9}
and
Y = {5,5,9,9,10,12,13}
Where their empirical CDFs are given as:
F_x(1) = 0.0833, F_x(3) = 0.25, F_x(5) = 0.33, F_x(7) = 0.5833 and F_x(9) = 1
and
F_y(5) = 0.2857, F_y(9) = 0.5714, F_y(10) = 0.7143, F_y(12) = 0.8571 and F_y(13) = 1
Assuming their joint distribution is
H(x,y) = F_x(x) * F_y(y)
which is actually the "assumption" of X and Y are independent.
How can i calculate the Z = X + Y and F(z) in MATLAB ?
Note: I gave the H(x,y) as a simple product function for the simplicity, but it can be anything in reality which actually models the dependency between X and Y.
Given continuous probability density functions FX and FY, with a joint probability density function FX,Y, we can compute FX+Y as the integral of FX,Y over the line z=x+y. If the probability density functions are discrete, the integral above should be written as the derivative of the integral over the part of the plane given by z<=x+y.
This is fairly simple to do in MATLAB. Let's start with OP's data:
F_x = [0.0833,0.0833,0.25,0.25,0.33,0.33,0.5833,0.5833,1]; % CDF
F_x = diff([0,F_x]); % PDF
F_y = [0,0,0,0,0.2857,0.2857,0.2857,0.2857,0.5714,0.7143,0.7143,0.8571,1]; % CDF
F_y = diff([0,F_y]); % PDF
H = F_x.' .* F_y; % example joint PDF
Now we sum F_cum(z) = sum(H(x,y)) for all values z<=x+y, and then take the derivative F = diff([0,F_cum]):
[m,n] = size(H);
F_cum = zeros(1,m+n-1);
for z = 1:numel(F_cum)
s = 0;
for x = 1:numel(F_x)
y = z-x+1;
y = max(min(y,n),1); % avoid out of bounds indexing
s = s + sum(H(x,1:y));
end
F_cum(z) = s;
end
F = diff([0,F_cum]);
Note that we defined y=z-x+1, meaning z=y+x-1. Thus F(1) corresponds to z=2. This is the lowest possible value that can come out of the sum of the two distributions, which we defined to start at 1.
The above can be simplified by padding H with zeros and shifting each row by one additional element. This lines up the line z=x+y on a column of the matrix, allowing us to use a trivial sum projection:
H = [H,zeros(m)];
for ii=2:m
H(ii,:) = circshift(H(ii,:),ii-1);
end
F_cum = cumsum(sum(H,1));
F_cum = F_cum(1:end-1); % last element we don't need
F2 = diff([0,F_cum]);
But because diff([0,cumsum(F)]) == F (up to numerical precision), we can skip those two operations:
F3 = sum(H,1);
F3 = F3(1:end-1); % last element we don't need
(all(abs(F-F2)<1e-15) and all(abs(F-F3)<1e-16))
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 1 year ago.
Improve this question
I was wondering if there is a way to determine if an image is blurry or not by analyzing the image data.
Another very simple way to estimate the sharpness of an image is to use a Laplace (or LoG) filter and simply pick the maximum value. Using a robust measure like a 99.9% quantile is probably better if you expect noise (i.e. picking the Nth-highest contrast instead of the highest contrast.) If you expect varying image brightness, you should also include a preprocessing step to normalize image brightness/contrast (e.g. histogram equalization).
I've implemented Simon's suggestion and this one in Mathematica, and tried it on a few test images:
The first test blurs the test images using a Gaussian filter with a varying kernel size, then calculates the FFT of the blurred image and takes the average of the 90% highest frequencies:
testFft[img_] := Table[
(
blurred = GaussianFilter[img, r];
fft = Fourier[ImageData[blurred]];
{w, h} = Dimensions[fft];
windowSize = Round[w/2.1];
Mean[Flatten[(Abs[
fft[[w/2 - windowSize ;; w/2 + windowSize,
h/2 - windowSize ;; h/2 + windowSize]]])]]
), {r, 0, 10, 0.5}]
Result in a logarithmic plot:
The 5 lines represent the 5 test images, the X axis represents the Gaussian filter radius. The graphs are decreasing, so the FFT is a good measure for sharpness.
This is the code for the "highest LoG" blurriness estimator: It simply applies an LoG filter and returns the brightest pixel in the filter result:
testLaplacian[img_] := Table[
(
blurred = GaussianFilter[img, r];
Max[Flatten[ImageData[LaplacianGaussianFilter[blurred, 1]]]];
), {r, 0, 10, 0.5}]
Result in a logarithmic plot:
The spread for the un-blurred images is a little better here (2.5 vs 3.3), mainly because this method only uses the strongest contrast in the image, while the FFT is essentially a mean over the whole image. The functions are also decreasing faster, so it might be easier to set a "blurry" threshold.
Yes, it is. Compute the Fast Fourier Transform and analyse the result. The Fourier transform tells you which frequencies are present in the image. If there is a low amount of high frequencies, then the image is blurry.
Defining the terms 'low' and 'high' is up to you.
Edit:
As stated in the comments, if you want a single float representing the blurryness of a given image, you have to work out a suitable metric.
nikie's answer provide such a metric. Convolve the image with a Laplacian kernel:
1
1 -4 1
1
And use a robust maximum metric on the output to get a number which you can use for thresholding. Try to avoid smoothing too much the images before computing the Laplacian, because you will only find out that a smoothed image is indeed blurry :-).
During some work with an auto-focus lens, I came across this very useful set of algorithms for detecting image focus. It's implemented in MATLAB, but most of the functions are quite easy to port to OpenCV with filter2D.
It's basically a survey implementation of many focus measurement algorithms. If you want to read the original papers, references to the authors of the algorithms are provided in the code. The 2012 paper by Pertuz, et al. Analysis of focus measure operators for shape from focus (SFF) gives a great review of all of these measure as well as their performance (both in terms of speed and accuracy as applied to SFF).
EDIT: Added MATLAB code just in case the link dies.
function FM = fmeasure(Image, Measure, ROI)
%This function measures the relative degree of focus of
%an image. It may be invoked as:
%
% FM = fmeasure(Image, Method, ROI)
%
%Where
% Image, is a grayscale image and FM is the computed
% focus value.
% Method, is the focus measure algorithm as a string.
% see 'operators.txt' for a list of focus
% measure methods.
% ROI, Image ROI as a rectangle [xo yo width heigth].
% if an empty argument is passed, the whole
% image is processed.
%
% Said Pertuz
% Abr/2010
if ~isempty(ROI)
Image = imcrop(Image, ROI);
end
WSize = 15; % Size of local window (only some operators)
switch upper(Measure)
case 'ACMO' % Absolute Central Moment (Shirvaikar2004)
if ~isinteger(Image), Image = im2uint8(Image);
end
FM = AcMomentum(Image);
case 'BREN' % Brenner's (Santos97)
[M N] = size(Image);
DH = Image;
DV = Image;
DH(1:M-2,:) = diff(Image,2,1);
DV(:,1:N-2) = diff(Image,2,2);
FM = max(DH, DV);
FM = FM.^2;
FM = mean2(FM);
case 'CONT' % Image contrast (Nanda2001)
ImContrast = inline('sum(abs(x(:)-x(5)))');
FM = nlfilter(Image, [3 3], ImContrast);
FM = mean2(FM);
case 'CURV' % Image Curvature (Helmli2001)
if ~isinteger(Image), Image = im2uint8(Image);
end
M1 = [-1 0 1;-1 0 1;-1 0 1];
M2 = [1 0 1;1 0 1;1 0 1];
P0 = imfilter(Image, M1, 'replicate', 'conv')/6;
P1 = imfilter(Image, M1', 'replicate', 'conv')/6;
P2 = 3*imfilter(Image, M2, 'replicate', 'conv')/10 ...
-imfilter(Image, M2', 'replicate', 'conv')/5;
P3 = -imfilter(Image, M2, 'replicate', 'conv')/5 ...
+3*imfilter(Image, M2, 'replicate', 'conv')/10;
FM = abs(P0) + abs(P1) + abs(P2) + abs(P3);
FM = mean2(FM);
case 'DCTE' % DCT energy ratio (Shen2006)
FM = nlfilter(Image, [8 8], #DctRatio);
FM = mean2(FM);
case 'DCTR' % DCT reduced energy ratio (Lee2009)
FM = nlfilter(Image, [8 8], #ReRatio);
FM = mean2(FM);
case 'GDER' % Gaussian derivative (Geusebroek2000)
N = floor(WSize/2);
sig = N/2.5;
[x,y] = meshgrid(-N:N, -N:N);
G = exp(-(x.^2+y.^2)/(2*sig^2))/(2*pi*sig);
Gx = -x.*G/(sig^2);Gx = Gx/sum(Gx(:));
Gy = -y.*G/(sig^2);Gy = Gy/sum(Gy(:));
Rx = imfilter(double(Image), Gx, 'conv', 'replicate');
Ry = imfilter(double(Image), Gy, 'conv', 'replicate');
FM = Rx.^2+Ry.^2;
FM = mean2(FM);
case 'GLVA' % Graylevel variance (Krotkov86)
FM = std2(Image);
case 'GLLV' %Graylevel local variance (Pech2000)
LVar = stdfilt(Image, ones(WSize,WSize)).^2;
FM = std2(LVar)^2;
case 'GLVN' % Normalized GLV (Santos97)
FM = std2(Image)^2/mean2(Image);
case 'GRAE' % Energy of gradient (Subbarao92a)
Ix = Image;
Iy = Image;
Iy(1:end-1,:) = diff(Image, 1, 1);
Ix(:,1:end-1) = diff(Image, 1, 2);
FM = Ix.^2 + Iy.^2;
FM = mean2(FM);
case 'GRAT' % Thresholded gradient (Snatos97)
Th = 0; %Threshold
Ix = Image;
Iy = Image;
Iy(1:end-1,:) = diff(Image, 1, 1);
Ix(:,1:end-1) = diff(Image, 1, 2);
FM = max(abs(Ix), abs(Iy));
FM(FM<Th)=0;
FM = sum(FM(:))/sum(sum(FM~=0));
case 'GRAS' % Squared gradient (Eskicioglu95)
Ix = diff(Image, 1, 2);
FM = Ix.^2;
FM = mean2(FM);
case 'HELM' %Helmli's mean method (Helmli2001)
MEANF = fspecial('average',[WSize WSize]);
U = imfilter(Image, MEANF, 'replicate');
R1 = U./Image;
R1(Image==0)=1;
index = (U>Image);
FM = 1./R1;
FM(index) = R1(index);
FM = mean2(FM);
case 'HISE' % Histogram entropy (Krotkov86)
FM = entropy(Image);
case 'HISR' % Histogram range (Firestone91)
FM = max(Image(:))-min(Image(:));
case 'LAPE' % Energy of laplacian (Subbarao92a)
LAP = fspecial('laplacian');
FM = imfilter(Image, LAP, 'replicate', 'conv');
FM = mean2(FM.^2);
case 'LAPM' % Modified Laplacian (Nayar89)
M = [-1 2 -1];
Lx = imfilter(Image, M, 'replicate', 'conv');
Ly = imfilter(Image, M', 'replicate', 'conv');
FM = abs(Lx) + abs(Ly);
FM = mean2(FM);
case 'LAPV' % Variance of laplacian (Pech2000)
LAP = fspecial('laplacian');
ILAP = imfilter(Image, LAP, 'replicate', 'conv');
FM = std2(ILAP)^2;
case 'LAPD' % Diagonal laplacian (Thelen2009)
M1 = [-1 2 -1];
M2 = [0 0 -1;0 2 0;-1 0 0]/sqrt(2);
M3 = [-1 0 0;0 2 0;0 0 -1]/sqrt(2);
F1 = imfilter(Image, M1, 'replicate', 'conv');
F2 = imfilter(Image, M2, 'replicate', 'conv');
F3 = imfilter(Image, M3, 'replicate', 'conv');
F4 = imfilter(Image, M1', 'replicate', 'conv');
FM = abs(F1) + abs(F2) + abs(F3) + abs(F4);
FM = mean2(FM);
case 'SFIL' %Steerable filters (Minhas2009)
% Angles = [0 45 90 135 180 225 270 315];
N = floor(WSize/2);
sig = N/2.5;
[x,y] = meshgrid(-N:N, -N:N);
G = exp(-(x.^2+y.^2)/(2*sig^2))/(2*pi*sig);
Gx = -x.*G/(sig^2);Gx = Gx/sum(Gx(:));
Gy = -y.*G/(sig^2);Gy = Gy/sum(Gy(:));
R(:,:,1) = imfilter(double(Image), Gx, 'conv', 'replicate');
R(:,:,2) = imfilter(double(Image), Gy, 'conv', 'replicate');
R(:,:,3) = cosd(45)*R(:,:,1)+sind(45)*R(:,:,2);
R(:,:,4) = cosd(135)*R(:,:,1)+sind(135)*R(:,:,2);
R(:,:,5) = cosd(180)*R(:,:,1)+sind(180)*R(:,:,2);
R(:,:,6) = cosd(225)*R(:,:,1)+sind(225)*R(:,:,2);
R(:,:,7) = cosd(270)*R(:,:,1)+sind(270)*R(:,:,2);
R(:,:,7) = cosd(315)*R(:,:,1)+sind(315)*R(:,:,2);
FM = max(R,[],3);
FM = mean2(FM);
case 'SFRQ' % Spatial frequency (Eskicioglu95)
Ix = Image;
Iy = Image;
Ix(:,1:end-1) = diff(Image, 1, 2);
Iy(1:end-1,:) = diff(Image, 1, 1);
FM = mean2(sqrt(double(Iy.^2+Ix.^2)));
case 'TENG'% Tenengrad (Krotkov86)
Sx = fspecial('sobel');
Gx = imfilter(double(Image), Sx, 'replicate', 'conv');
Gy = imfilter(double(Image), Sx', 'replicate', 'conv');
FM = Gx.^2 + Gy.^2;
FM = mean2(FM);
case 'TENV' % Tenengrad variance (Pech2000)
Sx = fspecial('sobel');
Gx = imfilter(double(Image), Sx, 'replicate', 'conv');
Gy = imfilter(double(Image), Sx', 'replicate', 'conv');
G = Gx.^2 + Gy.^2;
FM = std2(G)^2;
case 'VOLA' % Vollath's correlation (Santos97)
Image = double(Image);
I1 = Image; I1(1:end-1,:) = Image(2:end,:);
I2 = Image; I2(1:end-2,:) = Image(3:end,:);
Image = Image.*(I1-I2);
FM = mean2(Image);
case 'WAVS' %Sum of Wavelet coeffs (Yang2003)
[C,S] = wavedec2(Image, 1, 'db6');
H = wrcoef2('h', C, S, 'db6', 1);
V = wrcoef2('v', C, S, 'db6', 1);
D = wrcoef2('d', C, S, 'db6', 1);
FM = abs(H) + abs(V) + abs(D);
FM = mean2(FM);
case 'WAVV' %Variance of Wav...(Yang2003)
[C,S] = wavedec2(Image, 1, 'db6');
H = abs(wrcoef2('h', C, S, 'db6', 1));
V = abs(wrcoef2('v', C, S, 'db6', 1));
D = abs(wrcoef2('d', C, S, 'db6', 1));
FM = std2(H)^2+std2(V)+std2(D);
case 'WAVR'
[C,S] = wavedec2(Image, 3, 'db6');
H = abs(wrcoef2('h', C, S, 'db6', 1));
V = abs(wrcoef2('v', C, S, 'db6', 1));
D = abs(wrcoef2('d', C, S, 'db6', 1));
A1 = abs(wrcoef2('a', C, S, 'db6', 1));
A2 = abs(wrcoef2('a', C, S, 'db6', 2));
A3 = abs(wrcoef2('a', C, S, 'db6', 3));
A = A1 + A2 + A3;
WH = H.^2 + V.^2 + D.^2;
WH = mean2(WH);
WL = mean2(A);
FM = WH/WL;
otherwise
error('Unknown measure %s',upper(Measure))
end
end
%************************************************************************
function fm = AcMomentum(Image)
[M N] = size(Image);
Hist = imhist(Image)/(M*N);
Hist = abs((0:255)-255*mean2(Image))'.*Hist;
fm = sum(Hist);
end
%******************************************************************
function fm = DctRatio(M)
MT = dct2(M).^2;
fm = (sum(MT(:))-MT(1,1))/MT(1,1);
end
%************************************************************************
function fm = ReRatio(M)
M = dct2(M);
fm = (M(1,2)^2+M(1,3)^2+M(2,1)^2+M(2,2)^2+M(3,1)^2)/(M(1,1)^2);
end
%******************************************************************
A few examples of OpenCV versions:
// OpenCV port of 'LAPM' algorithm (Nayar89)
double modifiedLaplacian(const cv::Mat& src)
{
cv::Mat M = (Mat_<double>(3, 1) << -1, 2, -1);
cv::Mat G = cv::getGaussianKernel(3, -1, CV_64F);
cv::Mat Lx;
cv::sepFilter2D(src, Lx, CV_64F, M, G);
cv::Mat Ly;
cv::sepFilter2D(src, Ly, CV_64F, G, M);
cv::Mat FM = cv::abs(Lx) + cv::abs(Ly);
double focusMeasure = cv::mean(FM).val[0];
return focusMeasure;
}
// OpenCV port of 'LAPV' algorithm (Pech2000)
double varianceOfLaplacian(const cv::Mat& src)
{
cv::Mat lap;
cv::Laplacian(src, lap, CV_64F);
cv::Scalar mu, sigma;
cv::meanStdDev(lap, mu, sigma);
double focusMeasure = sigma.val[0]*sigma.val[0];
return focusMeasure;
}
// OpenCV port of 'TENG' algorithm (Krotkov86)
double tenengrad(const cv::Mat& src, int ksize)
{
cv::Mat Gx, Gy;
cv::Sobel(src, Gx, CV_64F, 1, 0, ksize);
cv::Sobel(src, Gy, CV_64F, 0, 1, ksize);
cv::Mat FM = Gx.mul(Gx) + Gy.mul(Gy);
double focusMeasure = cv::mean(FM).val[0];
return focusMeasure;
}
// OpenCV port of 'GLVN' algorithm (Santos97)
double normalizedGraylevelVariance(const cv::Mat& src)
{
cv::Scalar mu, sigma;
cv::meanStdDev(src, mu, sigma);
double focusMeasure = (sigma.val[0]*sigma.val[0]) / mu.val[0];
return focusMeasure;
}
No guarantees on whether or not these measures are the best choice for your problem, but if you track down the papers associated with these measures, they may give you more insight. Hope you find the code useful! I know I did.
Building off of Nike's answer. Its straightforward to implement the laplacian based method with opencv:
short GetSharpness(char* data, unsigned int width, unsigned int height)
{
// assumes that your image is already in planner yuv or 8 bit greyscale
IplImage* in = cvCreateImage(cvSize(width,height),IPL_DEPTH_8U,1);
IplImage* out = cvCreateImage(cvSize(width,height),IPL_DEPTH_16S,1);
memcpy(in->imageData,data,width*height);
// aperture size of 1 corresponds to the correct matrix
cvLaplace(in, out, 1);
short maxLap = -32767;
short* imgData = (short*)out->imageData;
for(int i =0;i<(out->imageSize/2);i++)
{
if(imgData[i] > maxLap) maxLap = imgData[i];
}
cvReleaseImage(&in);
cvReleaseImage(&out);
return maxLap;
}
Will return a short indicating the maximum sharpness detected, which based on my tests on real world samples, is a pretty good indicator of if a camera is in focus or not. Not surprisingly, normal values are scene dependent but much less so than the FFT method which has to high of a false positive rate to be useful in my application.
I came up with a totally different solution.
I needed to analyse video still frames to find the sharpest one in every (X) frames. This way, I would detect motion blur and/or out of focus images.
I ended up using Canny Edge detection and I got VERY VERY good results with almost every kind of video (with nikie's method, I had problems with digitalized VHS videos and heavy interlaced videos).
I optimized the performance by setting a region of interest (ROI) on the original image.
Using EmguCV :
//Convert image using Canny
using (Image<Gray, byte> imgCanny = imgOrig.Canny(225, 175))
{
//Count the number of pixel representing an edge
int nCountCanny = imgCanny.CountNonzero()[0];
//Compute a sharpness grade:
//< 1.5 = blurred, in movement
//de 1.5 à 6 = acceptable
//> 6 =stable, sharp
double dSharpness = (nCountCanny * 1000.0 / (imgCanny.Cols * imgCanny.Rows));
}
Thanks nikie for that great Laplace suggestion.
OpenCV docs pointed me in the same direction:
using python, cv2 (opencv 2.4.10), and numpy...
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
numpy.max(cv2.convertScaleAbs(cv2.Laplacian(gray, 3)))
result is between 0-255. I found anything over 200ish is very in focus, and by 100, it's noticeably blurry. the max never really gets much under 20 even if it's completely blurred.
One way which I'm currently using measures the spread of edges in the image. Look for this paper:
#ARTICLE{Marziliano04perceptualblur,
author = {Pina Marziliano and Frederic Dufaux and Stefan Winkler and Touradj Ebrahimi},
title = {Perceptual blur and ringing metrics: Application to JPEG2000,” Signal Process},
journal = {Image Commun},
year = {2004},
pages = {163--172} }
It's usually behind a paywall but I've seen some free copies around. Basically, they locate vertical edges in an image, and then measure how wide those edges are. Averaging the width gives the final blur estimation result for the image. Wider edges correspond to blurry images, and vice versa.
This problem belongs to the field of no-reference image quality estimation. If you look it up on Google Scholar, you'll get plenty of useful references.
EDIT
Here's a plot of the blur estimates obtained for the 5 images in nikie's post. Higher values correspond to greater blur. I used a fixed-size 11x11 Gaussian filter and varied the standard deviation (using imagemagick's convert command to obtain the blurred images).
If you compare images of different sizes, don't forget to normalize by the image width, since larger images will have wider edges.
Finally, a significant problem is distinguishing between artistic blur and undesired blur (caused by focus miss, compression, relative motion of the subject to the camera), but that is beyond simple approaches like this one. For an example of artistic blur, have a look at the Lenna image: Lenna's reflection in the mirror is blurry, but her face is perfectly in focus. This contributes to a higher blur estimate for the Lenna image.
Answers above elucidated many things, but I think it is useful to make a conceptual distinction.
What if you take a perfectly on-focus picture of a blurred image?
The blurring detection problem is only well posed when you have a reference. If you need to design, e.g., an auto-focus system, you compare a sequence of images taken with different degrees of blurring, or smoothing, and you try to find the point of minimum blurring within this set. I other words you need to cross reference the various images using one of the techniques illustrated above (basically--with various possible levels of refinement in the approach--looking for the one image with the highest high-frequency content).
I tried solution based on Laplacian filter from this post. It didn't help me. So, I tried the solution from this post and it was good for my case (but is slow):
import cv2
image = cv2.imread("test.jpeg")
height, width = image.shape[:2]
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
def px(x, y):
return int(gray[y, x])
sum = 0
for x in range(width-1):
for y in range(height):
sum += abs(px(x, y) - px(x+1, y))
Less blurred image has maximum sum value!
You can also tune speed and accuracy by changing step, e.g.
this part
for x in range(width - 1):
you can replace with this one
for x in range(0, width - 1, 10):
Matlab code of two methods that have been published in highly regarded journals (IEEE Transactions on Image Processing) are available here: https://ivulab.asu.edu/software
check the CPBDM and JNBM algorithms. If you check the code it's not very hard to be ported and incidentally it is based on the Marzialiano's method as basic feature.
i implemented it use fft in matlab and check histogram of the fft compute mean and std but also fit function can be done
fa = abs(fftshift(fft(sharp_img)));
fb = abs(fftshift(fft(blured_img)));
f1=20*log10(0.001+fa);
f2=20*log10(0.001+fb);
figure,imagesc(f1);title('org')
figure,imagesc(f2);title('blur')
figure,hist(f1(:),100);title('org')
figure,hist(f2(:),100);title('blur')
mf1=mean(f1(:));
mf2=mean(f2(:));
mfd1=median(f1(:));
mfd2=median(f2(:));
sf1=std(f1(:));
sf2=std(f2(:));
That's what I do in Opencv to detect focus quality in a region:
Mat grad;
int scale = 1;
int delta = 0;
int ddepth = CV_8U;
Mat grad_x, grad_y;
Mat abs_grad_x, abs_grad_y;
/// Gradient X
Sobel(matFromSensor, grad_x, ddepth, 1, 0, 3, scale, delta, BORDER_DEFAULT);
/// Gradient Y
Sobel(matFromSensor, grad_y, ddepth, 0, 1, 3, scale, delta, BORDER_DEFAULT);
convertScaleAbs(grad_x, abs_grad_x);
convertScaleAbs(grad_y, abs_grad_y);
addWeighted(abs_grad_x, 0.5, abs_grad_y, 0.5, 0, grad);
cv::Scalar mu, sigma;
cv::meanStdDev(grad, /* mean */ mu, /*stdev*/ sigma);
focusMeasure = mu.val[0] * mu.val[0];
Normally, a Gabor filter, as its name suggests, is used to filter an image and extract everything that it is oriented in the same direction of the filtering.
In this question, you can see more efficient code than written in this Link
Assume 16 scales of Filters at 4 orientations, so we get 64 gabor filters.
scales=[7:2:37], 7x7 to 37x37 in steps of two pixels, so we have 7x7, 9x9, 11x11, 13x13, 15x15, 17x17, 19x19, 21x21, 23x23, 25x25, 27x27, 29x29, 31x31, 33x33, 35x35 and 37x37.
directions=[0, pi/4, pi/2, 3pi/4].
The equation of Gabor filter in the spatial domain is:
The equation of Gabor filter in the frequency domain is:
In the spatial domain:
function [fSiz,filters,c1OL,numSimpleFilters] = init_gabor(rot, RF_siz)
image=imread('xxx.jpg');
image_gray=rgb2gray(image);
image_gray=imresize(image_gray, [100 100]);
image_double=double(image_gray);
rot = [0 45 90 135]; % we have four orientations
RF_siz = [7:2:37]; %we get 16 scales (7x7 to 37x37 in steps of two pixels)
minFS = 7; % the minimum receptive field
maxFS = 37; % the maximum receptive field
sigma = 0.0036*RF_siz.^2 + 0.35*RF_siz + 0.18; %define the equation of effective width
lambda = sigma/0.8; % it the equation of wavelength (lambda)
G = 0.3; % spatial aspect ratio: 0.23 < gamma < 0.92
numFilterSizes = length(RF_siz); % we get 16
numSimpleFilters = length(rot); % we get 4
numFilters = numFilterSizes*numSimpleFilters; % we get 16x4 = 64 filters
fSiz = zeros(numFilters,1); % It is a vector of size numFilters where each cell contains the size of the filter (7,7,7,7,9,9,9,9,11,11,11,11,......,37,37,37,37)
filters = zeros(max(RF_siz)^2,numFilters); % Matrix of Gabor filters of size %max_fSiz x num_filters, where max_fSiz is the length of the largest filter and num_filters the total number of filters. Column j of filters matrix contains a n_jxn_j filter (reshaped as a column vector and padded with zeros).
for k = 1:numFilterSizes
for r = 1:numSimpleFilters
theta = rot(r)*pi/180; % so we get 0, pi/4, pi/2, 3pi/4
filtSize = RF_siz(k);
center = ceil(filtSize/2);
filtSizeL = center-1;
filtSizeR = filtSize-filtSizeL-1;
sigmaq = sigma(k)^2;
for i = -filtSizeL:filtSizeR
for j = -filtSizeL:filtSizeR
if ( sqrt(i^2+j^2)>filtSize/2 )
E = 0;
else
x = i*cos(theta) - j*sin(theta);
y = i*sin(theta) + j*cos(theta);
E = exp(-(x^2+G^2*y^2)/(2*sigmaq))*cos(2*pi*x/lambda(k));
end
f(j+center,i+center) = E;
end
end
f = f - mean(mean(f));
f = f ./ sqrt(sum(sum(f.^2)));
p = numSimpleFilters*(k-1) + r;
filters(1:filtSize^2,p)=reshape(f,filtSize^2,1);
fSiz(p)=filtSize;
end
end
% Rebuild all filters (of all sizes)
nFilts = length(fSiz);
for i = 1:nFilts
sqfilter{i} = reshape(filters(1:(fSiz(i)^2),i),fSiz(i),fSiz(i));
%if you will use conv2 to convolve an image with this gabor, so you should also add the equation below. But if you will use imfilter instead of conv2, so do not add the equation below.
sqfilter{i} = sqfilter{i}(end:-1:1,end:-1:1); %flip in order to use conv2 instead of imfilter (%bug_fix 6/28/2007);
convv=imfilter(image_double, sqfilter{i}, 'same', 'conv');
figure;
imagesc(convv);
colormap(gray);
end
phi = ij*pi/4; % ij = 0, 1, 2, 3
theta = 3;
sigma = 0.65*theta;
filterSize = 7; % 7:2:37
G = zeros(filterSize);
for i=(0:filterSize-1)/filterSize
for j=(0:filterSize-1)/filterSize
xprime= j*cos(phi);
yprime= i*sin(phi);
K = exp(2*pi*theta*sqrt(-1)*(xprime+ yprime));
G(round((i+1)*filterSize),round((j+1)*filterSize)) =...
exp(-(i^2+j^2)/(sigma^2))*K;
end
end
As of R2015b release of the Image Processing Toolbox, you can create a Gabor filter bank using the gabor function in the image processing toolbox, and you can apply it to an image using imgaborfilt.
In the frequency domain:
sigma_u=1/2*pi*sigmaq;
sigma_v=1/2*pi*sigmaq;
u0=2*pi*cos(theta)*lambda(k);
v0=2*pi*sin(theta)*lambda(k);
for u = -filtSizeL:filtSizeR
for v = -filtSizeL:filtSizeR
if ( sqrt(u^2+v^2)>filtSize/2 )
E = 0;
else
v_theta = u*cos(theta) - v*sin(theta);
u_theta = u*sin(theta) + v*cos(theta);
E=(1/2*pi*sigma_u*sigma_v)*((exp((-1/2)*(((u_theta-u0)^2/sigma_u^2))+((v_theta-v0)^2/sigma_v^2))) + (exp((-1/2)*(((u_theta+u0)^2/sigma_u^2))+((v_theta+v0)^2/sigma_v^2))));
end
f(v+center,u+center) = E;
end
end