Performance Enhancement of Corrupted Images Using Independent Component Analysis by Kurtosis for Blind Source Separation After Curvelet Denoising

Abstract

The current paper proposes to remedy the problem of corrupted images using efficient digital signal processing (DSP) scheme. The proposed scheme named Curvelet denoising followed by Blind Source Separation (BSS) by Wavelet Packets Decomposition (WPD) and Kurtosis standard. The proposed scheme aims to enhance the performance of removing the noise from corrupted images. Two source images for Lena and Boat have been chosen for testing the simulations. The source images have been corrupted by sources of noise. A comparison study has been performed between the performance of the last studies and the proposed scheme. The outcomes of the results have confirmed that; of all the addressed techniques, the proposed technique is the most efficient for improving the picture quality.