Adaptive Scale Adjustment Design of Unsharp Masking Filters for Image Contrast Enhancement

Adaptive Scale Adjustment Design of Unsharp Masking Filters for Image Contrast Enhancement

The unsharp masking filter (UMF) has been widely used in image
processing front ends for contrast enhancement. The filter,
being easy to implement, is based on the concept of augmenting
a scaled and high-passed version of the image to itself. The UMF
performance is critically dependent on the generation of the highpassed
signal to be added as well as its associated scale factor.
However, the optimal choice of filter parameters still remains a
challenging task due to possible intensity clipping problems where
the filtered pixel magnitude is vulnerable to be out of the permitted
display ranges. In this research, an adaptive scheme is formulated
such that the scale is derived from the pixel intensity of
the input image. Specifically, pixels in the mid-range intensity will
be assigned a larger scaling factor according to a Gaussian-like
profile. In addition, the optimal profile coefficients and the width
of the high-pass generator window are determined by adopting
the particle swarm optimization algorithm.