The difficulty of imaging around metal implants has been overcome using pulse sequences such as MAVRIC SL and SEMAC. While the images are largely artifact-free, subtle intensity fluctuations, or ‘pile up’ artifacts still remain. The post-processing approach presented here uses information extracted from a spectral-domain model to identify and correct local fluctuations in image intensity. In these regions, the spectral bins are expanded using a moving average filter. When the expanded bins are combined, the pile up is significantly reduced. This pile up correction can improve the aesthetic quality of the images, and significantly improve their diagnostic ability.