A Remark On Relaxation of Integral Functionals