I am trying to plot clusters via the MATLAB function kmean but am getting way too many centroids and have no idea why. Here is my code and an example of a figure:
rng(1);
wv_prop = [min_pts(:) slope(:)];
if (isempty(wv_prop)==0)
[idx,C] = kmeans(wv_prop,2);
subplot(3,2,5);
plot(wv_prop(idx==1,1),wv_prop(idx==1,2),'b.','MarkerSize',12);
hold on
plot(wv_prop(idx==2,1),wv_prop(idx==2,2),'r.','MarkerSize',12);
plot(C(:,1),C(:,2),'kx',...
'MarkerSize',15,'LineWidth',3)
Here is an example of the data I use:
wv_prop:
-7.50904246127179e-05 2.52737793199461e-05
-7.64715493632322e-05 -29.2845021783221
-8.16630514296111e-05 -15.5896244315076
-8.60516901697005e-05 3.87325886247646e-05
-9.07390060961131e-05 4.06844795948271e-05
-7.93980060844007e-05 3.72806601486833e-05
-8.08420950480078e-05 3.81372062193057e-05
-8.53045358845788e-05 4.00072285969318e-05
-7.07712622172574e-05 3.55502071296987e-05
-8.02846575361635e-05 3.91085777803079e-05
-8.82904795076420e-05 4.21557386394776e-05
-8.32088783242009e-05 4.08103587885502e-05
-8.17564769131708e-05 4.06201592898485e-05
-8.88574631122910e-05 4.31980154605407e-05
-9.55496137235401e-05 4.55119867638717e-05
-7.11241881995855e-05 3.72772062250438e-05
-8.20641318582800e-05 6.09118479264444e-05
-7.92369664739745e-05 5.86246041439769e-05
-7.61219361068837e-05 5.57318660221894e-05
-8.52918510230295e-05 5.84710267850959e-05
-8.99668387994064e-05 5.84558301867090e-05
-9.62926333243702e-05 5.87762601336998e-05
-7.87678776488358e-05 4.67111894400931e-05
-7.53525297201741e-05 4.13207831828739e-05
-7.71766983561651e-05 3.82625914011195e-05
-9.03499693359608e-05 4.06874790212135e-05
-7.59387077492098e-05 2.92390401569819e-05
-7.97649576465785e-05 32.1683359898974
-8.06408560217508e-05 1.55409105433306e-05
-8.10515208048491e-05 1.31180389653758e-05
-7.70540121076476e-05 9.43353748786386e-06
-7.24001267378072e-05 5.78599898248438e-06
-8.93350436455590e-05 9.61034087028361e-06
-7.97722332494743e-05 4.89104076311932e-06
-8.40022599007737e-05 5.06726288587479e-06
-7.89655937936233e-05 2.44686642783556e-06
-8.58007004774045e-05 4.06628163987085e-06
-7.68775819259902e-05 1.06146142996962e-06
-7.05769224846652e-05 -2.25666633700963e-06
-7.73022200637920e-05 1.34546072255262e-06
-7.65784897728499e-05 1.62917829786978e-06
-7.41548367397790e-05 1.46536230997079e-06
-9.17371298592096e-05 1.17025036839378e-05
-7.35354500231489e-05 4.43710161064086e-06
function [] = Select_Figs(filename,startblock,endblock,startclust,endclust,animal,day)
%Select_Figs - Plots average waveforms, standard deviation, difference over time,
%fitted peak location histogram, mean squared error, k-mean clustered peak location and slope,
%and raw waveforms across selected blocks and clusters,
%saves to folder Selected-Figures-animal-date
%
%Select_Figs(filename,startblock,endblock,startclust,endclust,animal,date)
%
%filename - Sort.mat(e.g. = 'Sort.mat')
%
%startblock- first block (e.g. = 7)
%
%endblock - last block (e.g. = 12)
%
%startclust - first cluster (e.g. = 5)
%
%endclust - last cluster (e.g. = 10)
%
%animal - animal number (e.g. = 12)
%
%date - start of experiment (e.g. = 101617)
%
%Function called by User_Sort.m
Sort = filename;
addpath(pwd);
%Get Sort file
foldername = sprintf('Selected-Figures-%s-%s',animal,day); %Creates dynamic folder name to store figures
mkdir(foldername); %Makes directory
cd(fullfile(foldername)); %Cd to new directory
tvec = 0:.013653333:(.013653333*97); %Time vector
t = tvec(2:end);
for clust = startclust:endclust %Loops through all clusters
fig = cell(1,endblock); %Preallocate # of figures
name = sprintf('Idx_%d',clust); %Individual cluster name
fig{clust} = figure('Visible', 'off'); %Turns figure visibility off
for block = startblock:endblock %Loop through all blocks
wvfrms_avg =Sort.(name)(block).avg;
wvfrms_avg_scaled = (wvfrms_avg*10^6);
wvfrms_std =Sort.(name)(block).standdev;
min_ind = wvfrms_avg_scaled == min(wvfrms_avg_scaled);
min_loc = t(min_ind);
[~,io] = findpeaks(wvfrms_avg_scaled);
leftmin = io<find(wvfrms_avg_scaled==min(wvfrms_avg_scaled));
leftmin = leftmin(leftmin~=0);
rightmin = io>find(wvfrms_avg_scaled==min(wvfrms_avg_scaled));
rightmin = rightmin(rightmin~=0);
if (isempty(wvfrms_avg_scaled)==0)
subplot(3,2,1);
if (isnan(wvfrms_avg_scaled)==0)&((-30<min(wvfrms_avg_scaled))||(min_loc>0.55)||(min_loc<0.3)||(length(io(leftmin))>2)||(length(io(rightmin))>2))
plot(tvec(1:end-1),wvfrms_avg_scaled,'r');
else
plot(tvec(1:end-1),wvfrms_avg_scaled,'b');
end
end
new_wv = wvfrms_avg_scaled(40:end);
[~,locs_scaled] = findpeaks(new_wv);
if isempty(locs_scaled)==1
ind_scaled = max(new_wv);
else
ind_scaled = locs_scaled(1);
end
x1_scaled = new_wv(find(min(wvfrms_avg_scaled)));
y1_scaled = min(wvfrms_avg_scaled);
x2_scaled = ind_scaled;
y2_scaled = new_wv(find(ind_scaled));
slope_scaled= (y2_scaled-y1_scaled)./(x2_scaled-x1_scaled);
if (isnan(wvfrms_avg_scaled)==0)
if ((-30<min(wvfrms_avg_scaled)))
lab = sprintf('Time (ms) \n Peak exceeds amplitude range (%s)',num2str(min(wvfrms_avg_scaled)));
xlabel(lab,'FontSize',8);
ylabel('Mean Voltage (\muV)','FontSize',8);
title('Average Waveform','FontSize',8);
elseif ((min_loc>0.55)||(min_loc<0.3))
lab = sprintf('Time (ms) \n Peak location exceeds range (Time = %s)',num2str(min_loc));
xlabel(lab,'FontSize',8);
ylabel('Mean Voltage (\muV)','FontSize',8);
title('Average Waveform','FontSize',8);
elseif (length(io(leftmin))>2)||(length(io(rightmin))>2)
lab = sprintf('Time (ms) \n Peak limit exceeded (# = %s) Peak = %s',num2str(length(io)),num2str(min(wvfrms_avg_scaled)));
xlabel(lab,'FontSize',8);
ylabel('Mean Voltage (\muV)','FontSize',8);
title('Average Waveform','FontSize',8);
else
lab = sprintf('Time (ms) \n Peak = %s Slope = %s',num2str(min(wvfrms_avg_scaled)),num2str(slope_scaled));
xlabel(lab,'FontSize',8)
ylabel('Mean Voltage (\muV)','FontSize',8);
title('Average Waveform','FontSize',8);
end
end
if (isempty(wvfrms_std)==0&isempty(wvfrms_avg)==0)
subplot(3,2,2);
errorbar(t,wvfrms_avg,wvfrms_std); %Plots errorbars
end
wvfrms_num_text = sprintf(['Time (ms) \n # Waveforms: ' num2str(size(Sort.(name)(block).block,2))]);
xlabel(wvfrms_num_text,'FontSize',8);
ylabel('Mean Voltage (V)','FontSize',8);
title('Average Waveform + STD','FontSize',8);
wvfrms = Sort.(name)(block).block;
for i = 1:size(wvfrms,1)
if isempty(wvfrms)==0
min_pts = min(wvfrms,[],2); %Adds array of min wvfrm points to matrix
slope = zeros(1,size(wvfrms,1));
new = wvfrms(i,:);
new_cut = new(40:end);
[~,locs] = findpeaks(new_cut);
if isempty(locs)==1
ind = max(new_cut);
else
ind = locs(1);
end
x1 = new(find(min_pts(i)));
y1 = min_pts(i);
x2 = ind;
y2 = new(find(ind));
slope(i) = (y2-y1)./(x2-x1);
else
slope(i) = 0;
end
end
bins = 100;
hist_val = (min_pts(:)*10^6);
if isempty(hist_val)==0
%Convert matrix of min points to array and into microvolts
subplot(3,2,3);
histogram(hist_val,bins);
ylabel('Count','FontSize',8);
title('Waveform Peaks','FontSize',8);
cnt = histcounts(hist_val,bins); %Returns bin counts
line_fit = zeros(1,length(cnt)); %Preallocates vector to hold line to fit histogram
for i = 3:length(line_fit)-3
if (cnt(i)<mean(cnt)) %If bin count is less than mean, take mean of 3
cnt(i)=mean([cnt(i-1) cnt(i+1)]); %consecutive bins, set as bin count
end
if (mean([cnt(i-2) cnt(i-1) cnt(i) cnt(i+1) cnt(i+2)])>=mean(cnt)) %If mean of 5 consecutive bins
line_fit(i-1) = (max([cnt(i-2) cnt(i-1) cnt(i) cnt(i+1) cnt(i+2)]));%exceeds bin count, set max,
end %add to line fit vector
end
line_fit(line_fit<=mean(cnt)) = min(cnt)+1; %Set line_fit values less than mean
x = linspace(min(hist_val),max(hist_val),length(line_fit)); %X axis (min - max point of vals)
hold on
plot(x,line_fit,'k','LineWidth',1.5);
assignin('base','hist_val',hist_val);
if (isempty(hist_val)==0)
gm = fitgmdist(hist_val,2,'RegularizationValue',0.1);
warning('off','stats:gmdistribution:FailedToConverge');
comp1 = gm.ComponentProportion(1)*100;
comp2 = gm.ComponentProportion(2)*100;
mean1 = gm.mu(1);
mean2 = gm.mu(2);
hist_leg = sprintf('\\muV \n Component 1 = %0.2f%% Component 2 = %0.2f%% \n Mean 1 = %0.2f Mean 2 = %0.2f',comp1,comp2,mean1,mean2);
xlabel(hist_leg,'FontSize',8);
end
hold off
else
subplot(3,2,3);
hist_val = 0;
plot(hist_val);
end
hist_val = (slope(:)*10^3);
if isempty(hist_val)==0
subplot(3,2,4);
histogram(hist_val,bins);
ylabel('Count');
cnt = histcounts(hist_val,bins); %Returns bin counts
line_fit = zeros(1,length(cnt)); %Preallocates vector to hold line to fit histogram
for i = 3:length(line_fit)-3
if (cnt(i)<mean(cnt)) %If bin count is less than mean, take mean of 3
cnt(i)=mean([cnt(i-1) cnt(i+1)]); %consecutive bins, set as bin count
end
if (mean([cnt(i-2) cnt(i-1) cnt(i) cnt(i+1) cnt(i+2)])>=mean(cnt)) %If mean of 5 consecutive bins
line_fit(i-1) = (max([cnt(i-2) cnt(i-1) cnt(i) cnt(i+1) cnt(i+2)])); %exceeds bin count, set max,
end %add to line fit vector
end
line_fit(line_fit<=mean(cnt)) = min(cnt)+1; %Set line_fit values less than mean
x = linspace(min(hist_val),max(hist_val),length(line_fit)); %X axis (min - max point of vals)
hold on
plot(x,line_fit,'k','LineWidth',1.5);
gm = fitgmdist(hist_val,2,'RegularizationValue',0.1);
warning('off','stats:gmdistribution:FailedToConverge');
comp1 = gm.ComponentProportion(1)*100;
comp2 = gm.ComponentProportion(2)*100;
mean1 = gm.mu(1);
mean2 = gm.mu(2);
title('Waveform Slope','FontSize',8);
hist_leg = sprintf('Slope (m) \n Component 1 = %0.2f%% Component 2 = %0.2f%% \n Mean 1 = %0.2f Mean 2 = %0.2f',comp1,comp2,mean1,mean2);
xlabel(hist_leg,'FontSize',8);
hold off
else
subplot(3,2,4);
hist_val = 0;
plot(hist_val);
end
rng(1);
wv_prop = [min_pts(:) slope(:)];
if (isempty(wv_prop)==0)
[idx,C] = kmeans(wv_prop,2);
subplot(3,2,5);
plot(wv_prop(idx==1,1),wv_prop(idx==1,2),'b.','MarkerSize',12);
hold on
plot(wv_prop(idx==2,1),wv_prop(idx==2,2),'r.','MarkerSize',12);
plot(C(:,1),C(:,2),'kx',...
'MarkerSize',15,'LineWidth',3)
title('Clustered Peak and Slope','FontSize',8);
fig_about = sprintf('BL%s - Cluster %s Block %s', animal,num2str(clust),num2str(block));
figtitle(fig_about);
else
subplot(3,2,5);
wv_prop = 0;
plot(wv_prop);
end
if isempty(wvfrms)==0
[vals] = align_wvs(wvfrms);
if (~isempty(vals))
subplot(3,2,6);
plot(t,vals);
title('Raw Waveforms','FontSize',8);
end
else
subplot(3,2,6);
w = 0;
plot(w);
end
print(fig{clust},['Cluster-' num2str(clust) ' Block-' num2str(block)],'-dpng');
end
end
disp('Done');
end
In the below code, each y is of 8 min duration. Once I generate the Powerr matrix I have what I Need but How do i do 3 hour averages? Should i do within the for loop or once I generate the Powerr matrix can i still be able to find out the 3 hourly averages of my data? thanks in advance.
fid = fopen([directory '001/listing.txt'],'r');
tline = fgetl(fid);
ii = 0;
while (tline ~= -1)
if (length(tline) == 0)
continue
end
ii = ii + 1;
file_names(ii,:)=tline; % create file names in matlab data worksapace
year(ii) = str2num(tline(11:14));
month(ii) = str2num(tline(15:16));
day(ii) = str2num(tline(17:18));
hour(ii) = str2num(tline(20:21));
min(ii) = str2num(tline(22:23));
sec(ii) = str2num(tline(24:25));
tline = fgetl(fid);
end
fclose(fid);
timestamp = datenum(year,month,day,hour,min,sec);
%--------window applied is Hanning window--------
nfft = 4096; % |
hanning_window = hanning(nfft); % |
%-------------------------------------------------
timestamp = datenum(year,month,day,hour,min,sec);
tic
profile on
warning('off','MATLAB:audiovideo:wavread:functionToBeRemoved');
neglect_the_first_few= 5;
lastfile = 20;
Powerr = zeros(2049,20);
% fr = zeros(2049,1);
count =[];
parfor count = neglect_the_first_few + 1 :lastfile
warning('off','MATLAB:audiovideo:wavread:functionToBeRemoved');
y= [];
% for inputfile = infiles'
fn = [directory '/001/' file_names(count,:)];
[y,fs,nbits] = wavread(fn,'native');
disp(['sampling freq is ',num2str(fs),' hz. length of y is ',num2str(size(y)),'filenumber is ',num2str(count)])
% end
y = y- mean(y);
Y = double(y).*(1.49365435131571e-07);%CONVERSION - volts per bit for 24bit channel
% clear y;
nooverlap =[];%number of overlapped segments set to zero on Nov20th 2017
[Poww,fr] = pwelch(Y,hanning_window,nooverlap,nfft,fs);%no overlap taken into account%
Powerr(:,count) = Poww;
freq(:,count)= fr;
end
I'm writing the code in Matlab to find interest point using DoG in the image.
Here is the main.m:
imTest1 = rgb2gray(imread('1.jpg'));
imTest1 = double(imTest1);
sigma = 0.6;
k = 5;
thresh = 3;
[x1,y1,r1] = DoG(k,sigma,thresh,imTest1);
%get the interest points and show it on the image with its scale
figure(1);
imshow(imTest1,[]), hold on, scatter(y1,x1,r1,'r');
And the function DoG is:
function [x,y,r] = DoG(k,sigma,thresh,imTest)
x = []; y = []; r = [];
%suppose 5 levels of gaussian blur
for i = 1:k
g{i} = fspecial('gaussian',size(imTest),i*sigma);
end
%so 4 levels of DoG
for i = 1:k-1
d{i} = imfilter(imTest,g{i+1}-g{i});
end
%compare the current pixel in the image to the surrounding pixels (26 points),if it is the maxima/minima, this pixel will be a interest point
for i = 2:k-2
for m = 2:size(imTest,1)-1
for n = 2:size(imTest,2)-1
id = 1;
compare = zeros(1,27);
for ii = i-1:i+1
for mm = m-1:m+1
for nn = n-1:n+1
compare(id) = d{ii}(mm,nn);
id = id+1;
end
end
end
compare_max = max(compare);
compare_min = min(compare);
if (compare_max == d{i}(m,n) || compare_min == d{i}(m,n))
if (compare_min < -thresh || compare_max > thresh)
x = [x;m];
y = [y;n];
r = [r;abs(d{i}(m,n))];
end
end
end
end
end
end
So there's a gaussian function and the sigma i set is 0.6. After running the code, I find the position is not correct and the scales looks almost the same for all interest points. I think my code should work but actually the result is not. Anybody know what's the problem?
I'm trying to compute the average of every pixel with just the left and right neighbors but at the end of my processing I get only a white image, I can't find where my error. Here's my code
imageIn = imread('Prueba.jpg');
imageIn = rgb2gray(imageIn);
imageOut = zeros(size(imageIn));
ny = size(imageIn, 1);
nx = size(imageIn, 2);
imshow(imageIn);
u = [];
v = [];
tic
for i = 1:ny
u = imageIn(i,:);
v = zeros(1, ny);
for k = 2:ny-1
v(k) = (uint32(u(k-1))+uint32(u(k))+uint32(u(k+1)))/3;
end
%Special cases first and last pixel
v(1) = (uint32(u(2))+uint32(u(1))+uint32(u(2)))/3;
v(ny) = (uint32(u(ny-1))+uint32(u(ny))+uint32(u(ny-1)))/3;
imageOut(i,:) = v;
end
toc
imshow(imageOut);
Any ideas?
Change the last line of your code to imagesc(imageOut) and you'll see that the image is not in fact white.
Your code is fine; the reason the image appears white using the imshow() function is because after applying your local average the range of pixel intensities is considerably smaller and the default scaling used by imshow() is insufficient to bring out the contrast of the image.
Read about the difference b/t imshow() and imagesc() and you'll see the confusion.
Why not just create a 2nd matrix which is a clone of the first, shift it over and then averate the two matrices?
imIn = imread('Prueba.jpg');
nx = size(d,1);
ny = size(d,2);
% Create temporary matrices padded with nan
tmp1 = [nan(ny,2), d];
tmp2 = [d, nan(ny,2)];
imOut = tmp1;
imOut(:,:,2) = tmp2;
% use nanmean so the mean is just the value of the 1 column
imOut = nanmean(imOut,3);
out = imOut(2:end-1,:);
Try to use this
imageIn = imread('Prueba.jpg');
imageIn = rgb2gray(imageIn);
imageOut = zeros(size(imageIn));
ny = size(imageIn, 1);
nx = size(imageIn, 2);
imshow(imageIn);
u = [];
v = [];
tic
for i = 1:ny
u = imageIn(i,:);
v = zeros(1, ny);
for k = 2:ny-1
v(k) = (uint32(u(k-1))+uint32(u(k))+uint32(u(k+1)))/3;
end
%Special cases first and last pixel
v(1) = (uint32(u(2))+uint32(u(1))+uint32(u(2)))/3;
v(ny) = (uint32(u(ny-1))+uint32(u(ny))+uint32(u(ny-1)))/3;
imageOut(i,:) = v;
end
toc
imshow(imageOut);