Realization of Fast Median Filter Algorithm Based on the Relevance between Windows

Abstract

Through thinking deeply on the algorithm of standard median filter, the algorithm is improved in order to make it faster. A fast filtering algorithm is proposed with the help of the relevance between filtering windows. The improved algorithm takes the indexes of the array storing the data of an image as media cleverly and takes the method combining the one-time total ordering with insertion sorting to reduce the time of comparison and movement greatly. The result of actual test indicates that using the improved median filter algorithm, the processing speed is promoted much on the premise of holding the effect of removing noise.