Divide an image iteratively

The images are classified as resampled or not by the SVM. If the label of the classifed image is resampled which is one I want to divide the image again and apply the SVM until the exact location of resampling is found and draw a rectangle around the resampled area. My code for saved SVM classifier and main script is given below and the other coddes are attched. How to do this ?

    function label = classifyImages(f1,f2,f3,f4)
    load ImageFeaturesPredict.csv

    t = ImageFeaturesPredict(1:end,:);

    CompactMdl = loadCompactModel('SVMImageData'); 
    label = predict(CompactMdl,t); 

    if label==0
    fprintf('Non-Resampled');
    end

    if label==1
     fprintf('Resampled');
    end
    end

mainScript.m

    close all;
    clear all;
    clc;

    %img = imread('Original\ucid\ucid00101.tif');
    %img = imread('Original\ucid\ucid00099.tif');
    img = imread('Original\ucid\ImageDataSet\ucid00426.tif');
    img = rgb2gray(img);
    img = double(img);

   [f1,f3,f4,f5] = extractFeaturesFunctionNew(img);
   label = classifyImages()

if true
  % code
end  if (label==1)
  disp('This is a resampled image');
  %resampled location finder 
  resamplelocator(img);

    else
      disp('This image is not resampled');
    end

if true
  % code
end

[Image Analyst]

Unfortunately it seems you forgot to attach your tif images, so no one is likely to do anything until you do.