New algorithm for transient suppression for images due to incomplete or partial boundary data

Any n-dimensional signal processing problem involves manipulation with the input data set with the region of support bounded at some (or all) edges. The resulting output sequence is the convolution of the geometrically bounded input data set and an impulse response of some filter, which not necessarily has the same region of support as the input data set. Thus the ambiguity due to unknown values of the input data set beyond its region of support results in indeterminacy of all (or some) of the output sequence values. The paper presents an approach for resolving the problem of indeterminacy by employing the technique of minimum response variation for reconstructing the values of the input data set beyond its original region of support.