I have attached a code for line and word segmentation.. Code works well for line segment ion but not for WORD. Please do help me out on this... It is used for Kannada handwritten document

% Read image
i=imgetfile();
input=imread(i);
figure, imshow(input), title('Original Image');

% Deskew a document
[ im, theta ] = imdeskew( input );
figure, imshow(im), title('Deskewed image');

if size(im,3)==3 % RGB image %%Convert to gray scale
    im=rgb2gray(im);
end

%threshold = graythresh(im); %%Convert to binary image
%im =~im2bw(im,threshold);
%im = medfilt2(im, [3 3]);
im = bwareaopen(im,60); %%Remove all object containing fewer than 60 pixels
                     %converting image to black and white
%I3 = bwmorph(~im, 'thin', inf); 
%figure %%Show image binary image
%imshow(im);
%title('INPUT IMAGE WITHOUT NOISE')

%find Area of Interest:
a=(im);
ijk=~a;
L=sum(ijk,2);
L=0;
measurements = regionprops(L, 'PixelIdxList','Area','BoundingBox');
% Get indexes of those regions that are >= 20 in length.
fiveLongRegions = find([measurements.Area] >= 5);
theIndexes = vertcat(measurements(fiveLongRegions).PixelIdxList);
c=a;
c(theIndexes,:)=0;
figure, imshow(c);
title('AREA OF INTEREST FOR LINE SEGMENT');

%Segmentation of line:
[x,y]=size(c);
mat1=sum(c,2);%sum the elements of bwrowise and save into column matrix mat1
mat2=y-mat1;%subtract each element of the sum matrix(mat1) from the width length(no. of columns)
mat3=mat2~=0;
mat4=diff(mat3);
index1=find(mat4);
[q,w]=size(index1);%size of index2 matrix is q*w
kap=1;
lam=1;
%iii=1;
%ii=1;
while kap<((q/2)+1);%number of loops=number of lines
    k=1;
    mat5=([]);
       for j=(index1(lam)+1):1:index1(lam+1)
         mat5(k,:)=a(j,:); %store the line segmented matrix
           k=k+1;
       end

              lam=lam+2;   
          kap=kap+1;

         figure, imshow(mat5);
         %imsave();

         %%WORDS SEGMENTATION:

          bwword=im2bw(mat5);%convert the image im into binary(black and white)
          aword=~bwword;

          bword=sum(aword,1);
          figure, imshow(aword);

          %%Measure lengths of all "0" regions.
          measurementsword = regionprops(bword, 'Area', 'PixelIdxList');

          % Get indexes of those regions that are >= 20 in length.
        fiveLongRegionsword = find([measurementsword.Area] >= 5);
          theIndexesword = vertcat(measurementsword(fiveLongRegionsword).PixelIdxList);
          cword=aword;
         cword(:,theIndexesword)=1;
         %figure, imshow(c);

          dword=~cword;
          %figure, imshow(dword);
          %title('AREA OF INTERESRT FOR WORD SEGMENTATION IMAGE');

          [xword,yword]=size(cword);
          mat1word=sum(cword,1);%sum the elements of bwrowise and save into column matrix mat1
          mat2word=xword-mat1word;%subtract each element of the sum matrix(mat1) from the width length(no. of columns)
          mat3word=mat2word~=0;
          mat4word=diff(mat3word);
          index1word=find(mat4word);
          [qword,wword]=size(index1word);%size of index2 matrix is q*w
        kapword=1;
          lamword=1;
          %%iii=1;
          %%ii=1;

           while kapword<((wword/2)+1);%number of loops=number of lines
      kword=1;
     mat5word=([]);
         for jword=(index1word(lamword)+1):1:index1word(lamword+1)
          mat5word(:,kword)=bwword(:,jword); %store the word segmented matrix
            kword=kword+1;
        end

             lamword=lamword+1;   
         kapword=kapword+1;

         figure, imshow(mat5word);
           end 

end
------------------------------------------------------------------------------------------------------

%deskew

function [ im, theta ] = imdeskew( src, max_angle, resolution, plotOn ) 

% 0. parameter settings
if nargin <= 1
    max_angle   = 15;
    resolution  = .5;
    plotOn      = 1;
elseif nargin <= 2
    resolution  = .5;
    plotOn      = 1;
elseif nargin <= 3
    plotOn      = 1;
else
    error('unsupported input format')
end
% input settings
if size( src, 3 ) > 1 
    display('warning: automatic converting color image to binary')
    gray = rgb2gray( src );
    src  = gray > graythresh( gray ) * 255;
else
    if ~islogical( src )
        display('warning: automatic converting grayscale image to binary')
        src = src > graythresh( src ) * 255;
    end
end
% 1. extract black text pixels for analysis
[ h, w ]    = size( src );
[ text_x, text_y ] = ind2sub(  [ h,w ], find( src(:) == 0 ) );
% 2. compute the information entroy of a projection profile
angles = -max_angle : resolution : max_angle;
cx  = h/2;
cy  = w/2;
len = size( text_x, 1 );
hist_to_prob = @( h ) ( h( h ~= 0 ) / len );
score = [];
for a = angles
    sin_a = sin( a / 180 * pi );
    cos_a = cos( a / 180 * pi );
    sx    = round( ( text_x - cx ) * cos_a + ( text_y -cy ) * sin_a + cx );
    freq  = hist( sx, unique(sx) );
    prob  = hist_to_prob( freq );
    entropy = -prob .* log( prob );
    score(end+1) = sum( entropy(:) );
end
% 3. generate output
[ val, min_idx ] = min( score );
theta = -angles( min_idx );
im = not( imrotate( not( src ), theta, 'loose' ) );
if plotOn
    figure,subplot(131),plot( -angles, score ), xlabel( 'deskew angle: degree' ), ylabel( 'projection profile entropy ' );
    hold on, line( [ -angles( min_idx ), -angles( min_idx )]' , [ min( score )-.1, max( score ) ]', 'color',[1,0,0] )
    hold on, text(  -angles( min_idx )-1, max( score )-.2 , 'optimal deskew angle' ),axis square
    subplot(132),imshow( imerode(src, ones(3) ) ); title( 'before deskew ');
    subplot(133),imshow( imerode(im, ones(3) ) ); title( 'after deskew' );
end

[LOVEPREET KAUR]

which algorithm is used for this skew detection techique??? and please provide some explantion of this code