Nonlinear Enhancement by Functional Mapping

Linear enhancement can be seen as a mapping of wavelet coefficients by a linear function Em(x) = Gmx. A direct extension of this is a nonlinear mapping function Em(x). The main challenges here are how to design a nonlinear function and how to best utilize multichannel information extracted from a dyadic wavelet framework to accomplish contrast enhancement.

4.1 Minimum Constraint for an Enhancement Function

A major concern in enhancement is to introduce no artifacts during processing (analysis) and reconstruction (synthesis). For the dyadic wavelet framework this means that we should not create new extrema in any of the channel outputs. This defines a minimum constraint on any enhancement function: A function must be continuous and monotonically increasing.

4.2 Filter Selection

0 0

Post a comment