An algorithm and a pipelined architecture for order-statistic determination and L-filtering

An algorithm is presented for determination of order statistics, based on iterative adaptive thresholding of binary input data on a bit-row basis. It is shown that this algorithm lends itself very well to pipelined implementation and that a parallel pipelined device for L-filtering can be constructed from it, using the idea of a distributed arithmetic FIR (finite impulse response) digital filter. Such a device would be quite useful in image processing, since it would enable an operation which combines attributes of linear filtering with suppression of impulsive noise to be applied to data windows of useful size in real time.<>