Fpga implementation of Combinatorial filtering approach for image Denoising using qca technology

Abstract

Digital image processing based integrated circuits have gained much newlineimportance in recent years due to its wide range of applications in industry, newlineagriculture, remote sensing, and medical field. Developing those chips with newlineeffective utilization of resources with low area, lesser power and delay is a newlinechallenging task. Most of the works are focused on reducing the power and area for newlineimplementing the effective noise removal filtering. As noise is the dynamic nature newlinein the images with random distribution, the existing filtering techniques consumes newlinehigh area, more power and delay in chip level developments. newlinePrimarily, this work introducing the non-local means with packing of multi newlinepatches (NLM-PMP) with non-subsampled contourlet (NSC) approach. It is an newlineeffective method, which functions based on multi-scale directional and newlinedecomposition properties. Further improve the performance of the system, this newlinework also utilizes Gabor filter with image statistics (SGIF) approach for extra newlinesmoothing in the texture region by eliminating the ringing artifacts. This method newlineperfectly restores the small details, edges and texture information of image. The newlineperformance of proposed method is verified on the noisy greyscale images and it newlineeliminated the all low level miniature noises. The simulation results shows that the newlineproposed method resulted in better subjective and objective performance compared newlinestate of art approaches for various values of standard deviation. However, this newlinework gives the better visual and quantitative results. As this work implemented newlineusing Matlab environment, the optimization of resource utilization is not possible. newlineThus, the next works are focused on FPGA based architecture. newlineVII newlineSecondly, the research majorly focuses on hardware implementation of hybrid newlinedecision-based window mask generation (HDWMG) based Gaussian filter (GF) to newlinesuppress the various sources of noise focusing both the frequency dependent noise newlineand random valued impulse noise. In addition, proposed HDWMG auto changed newlinethe standard noise deviation with