How can I reduce the time complexity of this matlab code.. Its taking lot of time to execute when an image is fed as input.

    cnt1=0; cnt2=0; cnt3=0; cnt4=0; cnt5=0; cnt6=0; cnt7=0; cnt8=0;
    p = 1;
    row = 1;

    while p<(size(Theta_deg_img,1))
        q = 1;   
        while q<(size(Theta_deg_img,2))
            for i=p:p+8             
                for j=q:q+8                                                 
                    if Theta_deg_img(i,j)>=-180 && Theta_deg_img(i,j)<=-135                    
                        cnt1 = cnt1+1;                
                    elseif Theta_deg_img(i,j)>=-134 && Theta_deg_img(i,j)<=-90                   
                        cnt2 = cnt2 + 1;                                    
                    elseif Theta_deg_img(i,j)>=-89 && Theta_deg_img(i,j)<=-45                    
                        cnt3 = cnt3 + 1;                
                    elseif Theta_deg_img(i,j)>=-44 && Theta_deg_img(i,j)<=0                    
                        cnt4 = cnt4 + 1;                
                    elseif Theta_deg_img(i,j)>=1 && Theta_deg_img(i,j)<=45                    
                        cnt5 = cnt5 + 1;                
                    elseif Theta_deg_img(i,j)>=46 && Theta_deg_img(i,j)<=90                   
                        cnt6 = cnt6 + 1;                
                    elseif Theta_deg_img(i,j)>=91 && Theta_deg_img(i,j)<=135                 
                        cnt7 = cnt7 + 1;                                
                    else                        
                        cnt8 = cnt8 + 1;
                    end                    
                    eval([fsrc '(row,1) = cnt1;']);
                    eval([fsrc '(row,2) = cnt2;']);
                    eval([fsrc '(row,3) = cnt3;']);
                    eval([fsrc '(row,4) = cnt4;']);
                    eval([fsrc '(row,5) = cnt5;']);
                    eval([fsrc '(row,6) = cnt6;']);
                    eval([fsrc '(row,7) = cnt7;']);
                    eval([fsrc '(row,8) = cnt8;']);                                     
                end                               
            end           
            q = q + 9;  row = row+1;        
            cnt1=0; cnt2=0; cnt3=0; cnt4=0; cnt5=0; cnt6=0; cnt7=0; cnt8=0;                    
        end         
        p = p + 9;        
    end

Answer 1

[Harshit]

Vectorize your code. Example you don't need inner loops you can use sum to get the value of counts. In the outer two loops you are calculating size again and again. Store it in a variable.

[.. Vamshi]

Thank you Harshit. That's true.. I should have stored size in a variable. How can I use sum for counts?

[Harshit]

 % i=p:p+8
   j = q:q+8
   count1 = count1+sum(Theta_deg_img(i,j)>=-180 && Theta_deg_img(i,j)<=-135)

[.. Vamshi]

mmm... I want here to increment the count every time when the condition is met. Does the above piece of code satisfy my requirement?

[Harshit]

Yes it will just think about it. Whenver the condn is true 1 will be generated in the matrix and otherwise 0. sum will give the no of 1.

[.. Vamshi]

I followed your suggestion.. the counts have to be stored for every iteration.. but here am unable to do..

Answer 2

[Akiva Gordon]

Also, try to avoid using eval. Look at this link to see why:

http://www.mathworks.com/matlabcentral/answers/51946-systematic-do-not-use-global-don-t-use-eval

[.. Vamshi]

Thank you A G. I think eval is fine here for concatenations, isn't it?