Image enhancement with nonlinear local interaction

A nonlinear lateral interaction is examined which adapts to local image characteristics. The models studied show an increased edge enhancement as the local mean increases and also show a gain adaptation (logarithmic transformation). The models have been formulated in a discrete signal processing framework and their behavior as digital image processing systems has been evaluated. The simplified model demonstrates all the major characteristics of interest for this class of model and allows for efficient exploration and experimentation of these characteristics as the model parameters change.<<ETX>>