Impulsive noise is a problem frequently occurred in image processing. This problem gets an importance especially when the important details in images having highly intensive impulsive noise are required to be retrieved. In this work, an artificial neural network method was developed for the purpose of reducing impulsive noise in images without removing the fine details. By using a back propagation algorithm, fine details in images with high noise intensity were obtained. Furthermore, the results were also compared with the traditional methods such as median filters. © 2005 IEEE.