Contrast Adjustment of Clustered Undecimated Wavelet Approximation Coefficients for Colour Image Enhancement

Colour image enhancement is an emerging research area in various fields like forensics, atmospheric science, astrophotography, oceanography and medical imaging. In this paper an efficient colour image enhancement technique is proposed based on Undecimated Wavelet Transform (UWT). In order to preserve the chromatic information, at first the given RGB image is converted into HSI (Hue, Intensity and Saturation) colour space and the intensity component alone is transformed by the UWT transform into lower and higher sub bands. Then clustering is taken place for lower band coefficient using K-means clustering (KMC) where number of cluster (K) is predicted by hill climbing algorithm. To enhance the image, each clustered partition is enhanced using histogram equalization (HE). Finally, the enhanced RGB colour image is obtained from the enhanced intensity component along with the original hue and saturation components. The performance is evaluated by Peak Signal to Noise Ratio (PSNR), Structural SIMilarity (SSIM) index, Universal Quality Index (UQI) and Entropy. Experimental results show that the proposed system enhances the given image with very good visual quality and achieves over 90% SSIM and UQI with very good PSNR.