why the code given below is used in the watermarking DCT?

code is i=[]; j=[]; w=1; wmrk=watermark; welem=numel(wmrk); % welem - no. of elements for k=1:4096 kx=(x{k}); % Extracting block into kx for processing for i=1:8 % To address row of block for j=1:8 % To adress column of block if (i==8) && (j==8) && (w<=welem) % Eligiblity condition to insert watremark % i=1 and j=1 - means embedding element in first bit of every block if wmrk(w)==0 kx(i,j)=kx(i,j)+35; elseif wmrk(w)==1 kx(i,j)=kx(i,j)-35; end end end end w=w+1; x{k}=kx; kx=[]; % Watermark value will be replaced in block end >> give me the explanation of the line kx(i,j)=kx(i,j)+35; please help, thanking you!