In this paper, we propose a method to detect and eliminate the streaking artifacts, present in Feldkamp-David-Kress (FDK) based 3D cone-beam tomography, by comparing and combining two synthetic data sets. The data sets, which are 3D reconstructed volume data, are created from the original and negative of the X-ray projections of 3D Sheep-Logan model. The 2nd data set, created from the negative of the projections, has minimized streaking artifacts originated from high intensity region. However, it loses the bone structure information, so the original data set, where the streaking artifact is evident, is combined with the 2nd data set to recover bone structure information. The simulation results show a significant reduction of streaking artifact is achieved.