A new strategy of image denoising using multiplier-less FIR filter designed with the aid of differential evolution algorithm

Due to the rapid development of one-dimensional signal processing in last few decades, it has spread out its wings in the field of multi-dimensional signal processing too. This has mainly been dominated by the proposition and implementation of robust algorithms which have focused on efficient storage and reliable transmission of digital images of various kinds. During the transmission through wired or wireless medium, digital images often encountered different types of channel noise which can significantly distort its appearance. As a matter of fact, filtering operation of digital images forms one of the most important tasks to be performed at the receiving end. In this paper, we have proposed a novel design strategy of two-dimensional (2-D) low-pass filter by means of a powerful evolutionary optimization technique called Differential Evolution (DE) algorithm. Mask coefficients of the proposed filter are constrained to assume values as sum of powers-of-two, thus making the filter hardware friendly. Experimental results have demonstrated the power of the algorithm in reducing the effect of Gaussian noise from digital image in terms of various performance parameters like peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), image enhancement factor (IEF) and image quality index (IQI) and so on. A number of test images have been taken into our consideration for the purpose of establishing our proposition. Simulation results have confirmed the superiority of the proposed DE-based filter over the conventional low-pass filtering method.