Recursive weighted multi-plateau histogram equalization for image enhancement

Abstract

Histogram equalization is broadly used for contrast enhancement. However, it does not take brightness preservation into account. It is not suitable for real time applications like medical, synthetics aperture radar (SAR), consumer electronic products, etc. whereas the brightness preservation is of vital importance to avoid the annoying artifacts. This paper proposes a histogram equalization based method called recursive weighted multi-plateau histogram equalization (RWMPHE). The essential idea is to segment the histogram into two or more sub histograms followed by clipping with six plateau limits. Weighting process using normalized power law modifies the clipped histogram to equalize each partition independently using histogram equalization. Recursive mean separate histogram equalization (RMSHE), recursive sub-image histogram equalization (RSIHE) and recursively separated and weighted histogram equalization (RSWHE) are methods that perform similar histogram segmentation but do not carry out clipping with weighting for histogram modification. In this paper, it is shown that RWMPHE preserve the brightness and enhances the contrast more precisely and accurately than existing methods.