CrazyEngineers
  • Help with code: image classification

    sherly

    Member

    Updated: Oct 26, 2024
    Views: 1.1K
    hi
    i m a final year BE student working on image processing
    i need help on my project
    i have divided a 512 x 512 image into 64 blocks
    i have extracted features (mean,standard deviation) from one of the 64 blocks
    i need to extend this for all the blocks
    the for loop is not working
    i also need to create a feature library for vistex database
    can anyone help pleaseeeeeeeeeee:sad:
    thanks in advance
    0
    Replies
Howdy guest!
Dear guest, you must be logged-in to participate on CrazyEngineers. We would love to have you as a member of our community. Consider creating an account or login.
Replies
  • Kaustubh Katdare

    AdministratorJan 26, 2009

    Re: image classification

    How about sharing the part of the code that is troubling you?

    PS: When you post the code, do not forget to wrap it using
     ... [/ code] tags :)
    Are you sure? This action cannot be undone.
    Cancel
  • sherly

    MemberJan 26, 2009

    code troubling me

    hi
    i m a final year BE student working on image processing
    i need help on the same
    i have acquired 512 x 512 image from a database and divided it into 64 blocks.Our algorithm requires us to genetrate
    co occurence histograms for eight angles.
    we have extracted features from one block.
    we need to extend it for other blocks also and store it in a database.
    the following is the code for one amgle(0 degree)
    X1=imread(strcat(y,t{i},'.bmp'));
    imshow(X1)
    end
    I2=X1(:,:,1);
    g=cell(size(64));
    for i=1:512
    for j=1:512
    C(i,j)=I2(i,j);
    end
    end
    g=mat2cell(C,64*ones(8,1),64*ones(1,8));
    F=g{3,8};
    for i=0:255
    gray(i+1)=i;
    end
    % decomposing an image using haar wavelet transform
    [cA1,cH1,cV1,cD1] = dwt2(F,'db1');
    A1 = upcoef2('a',cA1,'db1',1); 
    H1 = upcoef2('h',cH1,'db1',1);
    V1 = upcoef2('v',cV1,'db1',1);
    D1 = upcoef2('d',cD1,'db1',1);
     
     
    %decomposing complement of an image using haar wavelet transform
     
    FC=imcomplement(F);
    [cA2,cH2,cV2,cD2] = dwt2(FC,'db1');
    A2 = upcoef2('a',cA2,'db1',1); 
    H2 = upcoef2('h',cH2,'db1',1);
    V2 = upcoef2('v',cV2,'db1',1);
    D2 = upcoef2('d',cD2,'db1',1);
     
     
     
    %histogram for A1 and D11 (d=1,angle=0)
    a11=wcodemat(A1,192);
    d11=wcodemat(H1,192);
    for i=1:64
    for j=1:64
    s1=a11(i,j);
    if(j+1<=64)
    s2=a11(i,j+1);
    else
    s2=0;
    end
    v1=d11(i,j);
    if(j+1<=64)
    v2=d11(i,j+1);
    else
    v2=0;
    end
    min1=min(s1,v2);
    min2=min(s2,v1);
    alpha=max(min1,min2);
    if(alpha==min1)
    His1(i,j)=s1;
    else
    His1(i,j)=0;
    end
    if(alpha==min2)
    His2(i,j)=s1;
    else
    His2(i,j)=0;
    end
    end
    end
     
     
     
    %histogram1
    min_intensity=min(min(His1));
    max_intensity=max(max(His1));
    [r1 c1]=size(His1);
    [r2 c2]=size(His2);
     
    init=0;
    sum=0;
    for a=0:255
    count=0;
    for i=1:1:r1
    for j=1:1:c1
    if(His1(i,j)==a)
    count=count+1;
    end
     
    end
    end
    prob(a+1)=count/4096;
    cum(a+1)=prob(a+1)+init;
    init=cum(a+1);
    sum=sum+init;
    mean1=sum/256;
    end
    mean1
     
    std=0;
    for a=0:255
    temp=abs(cum(a+1)-mean1);
    std=std+temp;
    end
    std1=std/256
     
     
    figure,
    stem(gray,cum),title('A1 and D11 d=1,angle=0');
      
    help me please.........:sad:
    Are you sure? This action cannot be undone.
    Cancel
Home Channels Search Login Register