Please explain this QR code program to me.

I wanna someone explain to me what every functions do ,special functions QRkey and im2freqs ?total code will extract Qr code from image

function [ hit ] = QRkey( img,L,RP ) hit=[]; [row ~] = size(img); for i = 1:row freq = im2freqs(img(i,:)); [num ~] = size(freq); for j = 3:num-3 Error = 0.5; if abs((freq(j,2)/3)-freq(j-2,2)) > ceil((freq(j,2)/3)*Error)+1  ... abs((freq(j,2)/3)-freq(j-1,2)) > ceil((freq(j,2)/3)*Error)+1  ... abs((freq(j,2)/3)-freq(j+1,2)) > ceil((freq(j,2)/3)*Error)+1  ... abs((freq(j,2)/3)-freq(j+2,2)) > ceil((freq(j,2)/3)*Error)+1  ... freq(j,3)==0  L(i,freq(j,1))==0  L(i,freq(j+2,1))==0  L(i,freq(j-2,1))==0 continue; end %x = freq(j,1) , y = i Box = RP( L(i,freq(j,1)) ).BoundingBox; Cdif = uint32(sum(abs(RP(L(i,freq(j+2,1))).Centroid - RP(L(i,freq(j,1))).Centroid))); Ar = RP( L(i,freq(j,1)) ).Area;
            if  Cdif <= 5 && ...
            L(i,freq(j-2,1))==L(i,freq(j+2,1)) && L(i,freq(j-2,1))~=L(i,freq(j,1))
                hit = [hit; L(i,freq(j,1))];
                j = j+2;
            end
        end
    end
    end
    function [ freq ] = im2freqs( img)
 col = numel(img);
    freq = [];
    co = [1 1 img(1)]; % [center counter color]
    for j = 2:col
        if co(3) ~= img(j)
            co(1) = uint32(co(1))+ uint32(co(2))/2;
            freq = [freq ; co];
            co = [j 1 img(j)];
        else
            co(2) = co(2) + 1;
        end
    end
    co(1) = uint32(co(1))+ uint32(co(2))/2;
    freq = [freq;co];
end
function QR = CropQR (img, points) %u = [points(2,:)-points(4,:) 0]; %v = [1 0 0]; %c = cross(u,v); %angle = sign(c(3))*180/pi*atan2(norm(c),dot(u,v)); %angle = -(mod(angle +360,360)-135); angle = atand((points(2,2)-points(4,2))/(points(2,1)-points(4,1))); angle = angle - 45; if points(4,1) < points(2,1) angle = angle + 180; end
   [h ,w ,~] = size(img);
   QR = imrotate(img, angle);
   [rh ,rw ,~] = size(QR);
   T=[w/2 h/2];
   RT=[rw/2 rh/2];
   R=[cosd(angle) -sind(angle); sind(angle) cosd(angle)];

   T = [T;T;T;T];
   RT = [RT;RT;RT;RT];
   rotpoints=(points-T)*R+RT ;

   rotpoints=uint32(rotpoints);
   xymax = max(rotpoints);
   xymin = min(rotpoints);

   QR = QR(xymin(2):xymax(2), xymin(1):xymax(1), :);
end
function [ ] = QRextraction( Rimg ) Rimg = imread('6.1.bmp'); img = rgb2gray(Rimg);
    %sharppining the image
    %enhancment image % and convert to binary image (black and white)
    img = imsharpen(img,'Radius',10,'Amount',1);    
    img = ~im2bw(img,0.5);

    [L num] = bwlabel(img)
    %now we need properties for each 8 component
    RP = regionprops(L,'all');
    hit = QRkey(img,L,RP);
    hit = [unique(hit) histc(hit,unique(hit))]

    figure,imshow(L ); 

    flag = zeros((numel(hit)/2));
    for i = 1:(numel(hit)/2)-1
        if flag(i)==1
            continue;
        end
        ci = RP(hit(i,1)).Centroid;
        j = i+1;
        if flag(j)==1
            continue;
        end
        cj = RP(hit(j,1)).Centroid;
        for k = 1:(numel(hit)/2)
            if i == k || j == k || flag(k)==1 
                continue;
            end
            ck = RP(hit(k,1)).Centroid;
            %start
            a = pdist([ci ; cj],'euclidean');
            b = pdist([ci ; ck],'euclidean');
            c = pdist([cj ; ck],'euclidean');
            diff = cj - ci;
            if c<a || c<b
                continue;
            end
            d = ck+diff;
            zz = abs((a*a + b*b)-(c*c))/(c*c)
            center = ((cj+ci+ck+d)/4);
            if zz < 0.1 
                x = [cj(1);ci(1);ck(1);d(1)];
                y = [cj(2);ci(2);ck(2);d(2)];

                x = x+(x-center(1))*0.5;
                y = y+(y-center(2))*0.5;
                flag(i)=1;flag(j)=1;flag(k)=1;
                final_image=CropQR( Rimg , [x y]);
                figure,imshow(final_image);

                break;              
            end

        end
    end

[Walter Roberson]

This appears to be code from https://github.com/FCIS2020-Projects/Automatic-Segment-QR-Code . Have you consulted the author documentation ?

[Image Analyst]

We can't understand it either (or don't want to take the extensive time to do so).

I suggest you contact the author and ask them why they didn't put in any comments like all good programmers do. This code would get a grade of D or F from me. virtually no comments, cryptic variable names, multiple statements on one line, bad indenting, etc. It's a mess.

https://blogs.mathworks.com/loren/category/best-practice/?s_tid=Blog_loren_Category

[S. M]

yes sir ,he didn't put any comment or documents

[Adam Danz]

You shouldn't share other people's work without at least giving them credit or providing a link to the source.

[S. M]

Yes ,but Thier work in open source code ,i didn't take private code

[Walter Roberson]

The author of the code appears to be Osama Fawzy at Faculty of Computer and Information Science Ain Shams University

[S. M]

Okay sir

Answer 1

[yanqi liu]

yes，sir，i think the key is to judge the square corner，such as

now，we can find the rule

a = pdist([ci; cj],'euclidean');
b = pdist([ci; ck],'euclidean');
c = pdist([cj; ck],'euclidean');
zz = abs((a*a + b*b)-(c*c))/(c*c);

when zz < 0.1 ，the area is QR area，such as