Correlation factor of window function

Hello Everyone,

I want to ask a question. I am working on a project of Short term Fourier transform in which i am multiplying Gaussian window with an input image. the process i am doing using Gaussian window code by Qian Kemao, the output response of Gaussian window was image as like input image. But when i checked the correlation (corr2) between input image and output image... surprisingly the factor was just 0.189..which is very low, but i am not getting any scientific approach to this result beside this in code the value of output is changing on every iteration, as per coding.so the last value is just the updated version. Can anyone please mention any other reasons for this result. what are the possible causes of this result? The coding is mentioned below

        for wyt=-0.5:0.1:0.5
        for wxt=-0.5:0.1:0.5
            w=w0.*exp(j*wxt*x+j*wyt*y);
            w=fexpand(w,mm,nn); 
            Fw=fft2(w);  
            sf=ifft2(Ff.*Fw); 
            sf=sf(1+sx:m+sx,1+sy:n+sy); 
            sf=sf.*(abs(sf)>=3); 
            sf=fexpand(sf,mm,nn); 
            gtemp=ifft2(fft2(sf).*Fw); 
            g.filtered=g.filtered+gtemp(1+sx:m+sx,1+sy:n+sy); 
        end
    end

I will be very thankful. Thanks Regards