Color image denoising in wavelet domain using adaptive thresholding incorporating the Human Visual System model

Abstract

This paper utilizes the characteristics of the Human Visual System (HVS) for obtaining better results in denoising color images corrupted by Gaussian noise. Implementation of the Contrast Sensitivity Function (CSF) of the HVS is done in each of the sub-bands in the wavelet domain. In this paper, an 11-weight Invariant Single Factor (ISF) system is proposed and implemented which outperforms the generally used 6-weight ISF system. The major strength of the incorporation of this vision model is the optimum way in which the scales are distributed among the luminance and chrominance components to achieve better image quality without color degradation. Two adaptive thresholding schemes in the wavelet domain have been used in the denoising process. To evaluate the performance of the algorithm universal image quality index (Q) and Peak Signal-to-Noise Ratio (PSNR) have been used.