how to use logic array

i have a 8x6 logical array

result1 =

  8×6 logical array

   1   0   0   0   0   0
   0   0   0   1   0   0
   0   0   0   0   1   0
   0   0   0   0   0   0
   0   1   0   0   0   0
   0   0   1   0   0   0
   0   0   0   0   0   1
   0   0   0   0   0   0

how to use this logical array to extract the values in two matrix assosiated with logical ones

[madhan ravi]

You need to be clear how does your look like? the sizes??

[karishma koshy]

these are the two arrays:

  329.8325  361.1270
  613.6074  215.8294
  376.3128  310.3919
  636.6237  446.4107
  111.5548  191.3145
  248.6901  372.9032
  338.8193    3.1024
  428.1734  211.9613

row_frame2_adata =

  329.7856  364.9094
  111.9948  187.5294
  253.1863  368.9889
  614.6203  217.2036
  378.9097  307.9001
  259.9949   27.1064

the elements that nearly similar are linked and gives output as one. how can I use to find the elements linked . That is 1 is second column shows that 111.5548 191.3145 in frame 1 is linked to 111.9948 187.5294 in frame 2.

but i want to use the one in logical array to find all the elements linked similarly as mentioned above.

[KSSV]

What exactly is your problem?

[karishma koshy]

I want to find the values associated with the logical ones

[KALYAN ACHARJYA]

Madhan, already ask about sizes? Have you looked on it? I did not get any clue, just do combine of frame1 and frame2, give us the results with 8x6, rest is easy.

If you want to any logical operation between result1 and frame, both must be the same.

Or

Elaborate the question with simple example?

[karishma koshy]

I am trying to locate the values that linked to the logical ones. So that I can identify it's movement and add the ID number later on.

So all I want is a for loop where I can take the size of row_frame1_data array, use the assignment matrix that i get from the Munkres algorithm to find out which points in adjacent frames have been linked.

[Guillaume]

@karishma, you asked this question already and I asked for clarification. Instead of clarifying you deleted that question and reposted it with even less information.... and you get asked the same clarifications.

So,if you want help, rather than wasting everybody's time do answer the questions that you're asked and spend time explaining everything clearly.

The question that you deleted was a lot better expressed than what you've written above even though it wasn't clear. It's even less clear now.

Personally, I won't contribute anymore since my input wasn't thought worthwile.

[karishma koshy]

Hi Sir

I thought it was not clear and had less information. That why I resubmitted it added informations. Sorry for your inconvenience.

Answer 1

[Jon]

It isn't quite clear the specifics of what you want to retrieve but for example if you wanted to get the row in frame 1 that matches row 2 in frame 2 you could use

row = row_frame1_adata(result1(:,2),:)

[karishma koshy]

Hi Sir,

Yes, I want to retrieve values of frame 1 and frame 2 values associated with all logical ones. So that I can add ID number to those that are similar.

So all I want is a for loop where I can take the size of row_frame1_data array, use the assignment matrix that i get from the Munkres algorithm to find out which points in adjacent frames have been linked

[Jon]

Sorry, I'm still struggling to understand what exactly you are trying to do. I am thinking though that you have some ID's for the rows in frame1, say they are just the row numbers, 1,2,...8

Then the matching ID's for each row in frame2 would be given by

[id2,~] = find(result1)

If you have one more level of indirection, and your ids for frame 1 are given by a vector, call it id1, then you could get the ids in frame 2 using

[idx,~] = find(result1)
id2 = id1(idx)

Note it is important to use two left hand arguments (as I did above) when calling find so that it returns row and column indices rather than linear indices.

If I am still not understanding what you are trying to do, then please make a simple example which gives the inputs, and desired outputs that you are looking for.

[karishma koshy]

row_frame1_data =

  999.0000  329.8325  361.1270   22.0027
  999.0000  613.6074  215.8294   28.1989
  999.0000  376.3128  310.3919   22.9178
  999.0000  636.6237  446.4107   27.1672
  999.0000  111.5548  191.3145   24.8943
  999.0000  248.6901  372.9032   41.9425
  999.0000  338.8193    3.1024   23.3755
  999.0000  428.1734  211.9613   23.5272

  row_frame2_data =

   1.0e+03 *

    1.0000    0.3298    0.3649    0.0227
    1.0000    0.1120    0.1875    0.0236
    1.0000    0.2532    0.3690    0.0371
    1.0000    0.6146    0.2172    0.0307
    1.0000    0.3789    0.3079    0.0216
    1.0000    0.2600    0.0271    0.0244
    
    

  FRAME     CENTRE X  CENTRE Y   RADIUS   ID
 
  999.0000  329.8325  361.1270   22.0027   1
  999.0000  613.6074  215.8294   28.1989   4  
  999.0000  376.3128  310.3919   22.9178   5 
  999.0000  636.6237  446.4107   27.1672   
  999.0000  111.5548  191.3145   24.8943   2
  999.0000  248.6901  372.9032   41.9425   3
  999.0000  338.8193    3.1024   23.3755   6
  999.0000  428.1734  211.9613   23.5272
  1.0e+03 *

    1.0000    0.3298    0.3649    0.0227   1
    1.0000    0.1120    0.1875    0.0236   2
    1.0000    0.2532    0.3690    0.0371   3
    1.0000    0.6146    0.2172    0.0307   4
    1.0000    0.3789    0.3079    0.0216   5
    1.0000    0.2600    0.0271    0.0244   6

[karishma koshy]

My ultimate aim is to assign ID so i can track the moving object. Same as we label number in survillance camera to track the person's movement. here im trying to track multiple circles.

[Jon]

Here is a script which produces the results you show in your example

% example input arrays
row_frame1_data = [
  999.0000  329.8325  361.1270   22.0027
  999.0000  613.6074  215.8294   28.1989
  999.0000  376.3128  310.3919   22.9178
  999.0000  636.6237  446.4107   27.1672
  999.0000  111.5548  191.3145   24.8943
  999.0000  248.6901  372.9032   41.9425
  999.0000  338.8193    3.1024   23.3755
  999.0000  428.1734  211.9613   23.5272];

  row_frame2_data = 1e3*[
    1.0000    0.3298    0.3649    0.0227
    1.0000    0.1120    0.1875    0.0236
    1.0000    0.2532    0.3690    0.0371
    1.0000    0.6146    0.2172    0.0307
    1.0000    0.3789    0.3079    0.0216
    1.0000    0.2600    0.0271    0.0244];
    
    

result1 = logical(...
   [1   0   0   0   0   0
   0   0   0   1   0   0
   0   0   0   0   1   0
   0   0   0   0   0   0
   0   1   0   0   0   0
   0   0   1   0   0   0
   0   0   0   0   0   1
   0   0   0   0   0   0]);

% make arrays into tables
data1 = array2table(row_frame1_data,'VariableNames',{'frame','centre_x','centre_y','radius'});
data2 = array2table(row_frame2_data,'VariableNames',{'frame','centre_x','centre_y','radius'});

% assign id's to data2
data2.id = (1:size(data2,1))';

% find  row numbers in frame1 data that have  matching rows in frame2 data
% (these are the rows of result1 that have a 1 in them)
iRow = find(any(result1,2));

% find row numbers in frame 2 where each match occurs 
[id,~] = find(result1');

% assign these ids in data1
% rows that don't have a matching ID will be assigned NaN
data1.id = NaN(size(data1,1),1); 
data1.id(iRow) = id;

% display results
disp(data1)
disp(data2)

  

[karishma koshy]

Dear Sir

Thank you so much for your time and consideration sir,

That's exactly what I wanted.

[Jon]

Very good. Glad to hear it is working.

[karishma koshy]

Hi Sir, I have actually 1000 frames like this where I need ID number satisfying the above condition. Right now I'm getting only last two frames if I doing it in a loop.

I need ID number 1 for all the 1000 frames similar to what you have done.

How can I do that.

[Jon]

In each loop iteration you will have to assign your result to some variable (array, structure, table) that is indexed by the loop count. So, for example you could make an array of structures, let's call it S, with fields, frame, centre_x, centre_y, radius, id. At the end of each loop iteration you would then assign the results with something like:

% preallocate array of structures
numFrames = 1000
S(numFrames) = struct('frame',[],'centre_x',[],'centre_y',[],'radius',[],'id'[])
for iFrame = 1:numFrames
    % body of loop where you do all of the work
    % ...
    % save results in output structure
        S(iFrame).frame = data1.frame;
        S(iFrame).centre_x = data1.centre_x;
        S(iFrame).centre_y = data1.centre_y;
        S(iFrame).radius = data1.radius;
        S(iFrame).id = data1.id;
end

Note, that if you are doing all of the work inside of the loop, and all you care about is the final output structure, S, it is not really necessary to be creating those intermediate tables. I just used those in the example because it organized things well ,and gave a nicely formatted output with column headings. So if efficiency is a concern, you could rework the piece I gave you to do the same calculations but not bothering to put the values into the temporary tables. Alternatively, rather than using an array of structures, S, you could accumulate all of your results into a large table with an additional column indicating the frame count. You will have to work out the details for what works best for you.

[karishma koshy]

Hi Sir

i am new to matlab. Can I know how to accumulate all of my results into a large table with an additional column indicating the frame count.

ans =

  12×12 logical array

   1   0   0   0   0   0   0   0   0   0   0   0
   0   0   1   0   0   0   0   0   0   0   0   0
   0   0   0   0   0   0   0   0   0   0   1   0
   0   0   0   1   0   0   0   0   0   0   0   0
   0   0   0   0   0   0   0   0   0   1   0   0
   0   0   0   0   0   1   0   0   0   0   0   0
   0   1   0   0   0   0   0   0   0   0   0   0
   0   0   0   0   1   0   0   0   0   0   0   0
   0   0   0   0   0   0   0   0   1   0   0   0
   0   0   0   0   0   0   0   0   0   0   0   1
   0   0   0   0   0   0   0   1   0   0   0   0
   0   0   0   0   0   0   1   0   0   0   0   0


ans =

  12×10 logical array

   0   0   0   0   1   0   0   0   0   0
   0   0   1   0   0   0   0   0   0   0
   0   1   0   0   0   0   0   0   0   0
   0   0   0   0   0   0   0   1   0   0
   0   0   0   1   0   0   0   0   0   0
   0   0   0   0   0   1   0   0   0   0
   1   0   0   0   0   0   0   0   0   0
   0   0   0   0   0   0   0   0   0   1
   0   0   0   0   0   0   0   0   0   0
   0   0   0   0   0   0   0   0   0   0
   0   0   0   0   0   0   0   0   1   0
   0   0   0   0   0   0   1   0   0   0

    frame    centre_x    centre_y    radius    id
    _____    ________    ________    ______    __

      1        479.1      90.612     21.803     1
      1       653.78       337.1     25.102     2
      1       130.59      134.18     27.875     3
      1       297.99      323.38     39.806     4
      1       31.044      444.48     24.755     5
      1       350.96      53.437     43.816     6
      1       64.647      153.47     28.745     7
      1       293.48      169.92     36.579     8
      1       499.11       389.5     23.527     9
      1        426.5      328.29     22.962    10
      1        504.9      341.25     22.605    11
      1       620.52      413.07     28.185    12

    frame    centre_x    centre_y    radius    id
    _____    ________    ________    ______    __

      2       477.56      91.105     21.183     1
      2       64.259      150.29     27.273     7
      2       656.85      336.45      25.55     2
      2       297.95      324.68     40.698     4
      2       290.52      169.55     36.328     8
      2       344.63      48.536     43.103     6
      2       545.61      55.824     21.631     12
      2       354.06      380.63     25.841     11
      2       220.45      517.57     22.873     9
      2        139.3      281.79     27.241     5
      2       128.34      125.05     37.728     3
      2       244.67      359.59     28.057     10
      
      frame    centre_x    centre_y    radius    id
    _____    ________    ________    ______    __

      3       545.37      58.505     23.106     12
      3        657.4      334.62     26.254     2
      3       64.807      146.06     27.312     7 
      3       290.86      167.79     35.758     8
      3       477.65      89.518      21.65     1
      3       344.11      44.368     44.039     6
      3       621.32      408.27     24.244     13
      3       310.51      329.82     21.516     10
      3       72.627      113.72     23.659     3
      3       520.64       404.6      22.33     14
      
      
      
      
      

how can i get the above resultie the long table so on till 1000 frame

[Jon]

You could do something like:

% preallocate cell array of tables to collect output
numFrames = 1000
data = cell(numFrames,1)
for iFrame = 1:numFrames
    % body of loop where you do all of the work
    % ...
    % save results in output table
    data{iFrame} = data1;
end
% combine all of the individual tables into one large table
data = vertcat(data{:})

[karishma koshy]

In the body of loop can I use the previously sent code by you to assign ID number

[Jon]

I'm not completely clear on the details of your calculation but I think you should be able to adapt the code I sent you so that it can be included in the main loop. Since you seem to be doing a mapping of ids from the n+1 frame to the nth frame you probably have to be a little careful on the total loop count. Maybe it should only go to numFrames-1 so that you don't try to calculate beyond the end of your data

[karishma koshy]

Is there a way where I can do this way - for 1:size(number of points in frame i+1) -find assigned points in frame i -if no assigned points, assign the next available ID - if there is an assigned point, set ID for point in I+1 equal to ID of point in I+1

[Jon]

It sounds like you have a fairly clear understanding of what you want your code to do. You just need to translate your written description into MATLAB code that implements it. Just think carefully about each step you describe and I think you will be able to come up with the detailed MATLAB implementation. If you encounter some behavior that you really can't understand, or are stuck at some very specific point then ask for more help, but I think at this point you should be able to work out the details.

[karishma koshy]

Dear Sir,

I know what to do theoretically, but I'm new to coding.

That's why I am stuck with this.

[Jon]

Your application is quite complicated and probably is not a good place to start learning how to code. I would suggest that if you have not done so already you should first complete the free MATLAB On Ramp training,

https://matlabacademy.mathworks.com/

Make sure you work your way all the way through this and don't skip any steps.

After you complete the training and come back to your problem I think you will find you will be able to code a solution.

Answer 2

[Jackson Burns]

You can use Masking to index your array of results. Try this:

row_frame_data(result1)

Where row_frame_data is the 8*6 atrix of results and result1 is the logical matrix of the values you want.

[karishma koshy]

When I tried this I got an error saying the logical indices contain a true value outside the array bounds

[Mark Hayworth]

Exactly. Why is your logical matrix not the same size as the matrix you want to extract the values from. Your logical matrix has 4 more columns than your data. Why???

[karishma koshy]

In horizontal it's the number of elements in frame k and in vertical it's the number of elements in frame k+1 ie it's 8x6 array. The logical array was obtained using Hungarian algorithm

[Jackson Burns]

You will need to adjust the dimensions of your logical array to match the data.