A general framework for quadratic Volterra filters for edge enhancement

An inherent problem in most image enhancement schemes is the amplification of noise, which, due to Weber's law, is mostly visible in the darker portions of an image. Using a special class of quadratic Volterra filters, we can adapt the enhancement process in a computationally efficient way to the local image brightness because these filters are approximately equivalent to the product of a local mean estimator and a highpass filter. We analyze and derive this subclass of quadratic Volterra filters by investigating the 1-D case first, and then we generalize the results to two dimensions. An important property of these filters is that they map sinusoidal inputs to constant outputs, which allows us to develop a new filter characterization that is more intuitive for our application than the 4-D frequency response. This description finally leads to a novel least-squares design methodology. Image enhancement results using our Volterra filters are superior to those obtained with standard linear filters, which we demonstrate both quantitatively and qualitatively.