OBJECTIVE Positron emission tomography has been playing an important role as a quantitative molecular imaging modality to measure and image biochemical processes in vivo. The quality of positron emission tomography images may impact the accuracy of the quantitative or semiquantitative information extracted from them. METHODS In this study, the anisotropic diffusion filtering technique was used to de-noise emission images and attenuation maps of a whole body for improving the quantitative accuracy of emission images. The efficacy of the filtering technique was shown on a hot-lesion phantom as well as on whole-body emission images and attenuation maps. RESULTS The filtering technique showed a superb performance on images reconstructed using the iterative and the filtered back-projection reconstruction techniques. CONCLUSIONS Filtering may allow the reconstruction of images with optimal parameters with respect to lesion contrast and image resolution by removing the consequential noise. An improvement of as much as six-fold may be attained. It may also improve the accuracy of the quantitative information, such as standard uptake value extracted from emission images.