Genetic fuzzy optimized approximate multiplier design based non-linear anisotropic diffusion image denoising in VLSI

Abstract

Multiplier is a building block used in processor, embedded, Very Large-Scale Integration (VLSI) applications and integrated circuits. There are three main parameters of VLSI design lies in speed, area and power. Approximate computing is a developing one in digital design used to increase the performance. Many conventional methods are introduced for efficient approximate multiplier design. A new approximation multiplier design approach was presented with partial products of multiplier. A new approximate compressor was designed with higher power or speed for target precision. However, the designed approach failed to minimize the design complexity. In order to reduce power consumption and design complexity during the noise removal process, Genetic Fuzzy Optimized Approximate Multiplier based Non-Linear Anisotropic Diffusion Image Denoising (GFOAM-NLADID) Method is introduced. The key objective of the GFOAM-NLADID Method is to perform efficient approximate design using genetic fuzzy concepts for reducing the noise artifacts from input images. GFOAM-NLADID method used genetic fuzzy optimization algorithm to find optimal component (i.e., full adder type, compressor, and tree) for constructing the approximate multiplier design. The optimal component is selected after calculating the fuzzy fitness function. When the fitness function satisfies the fuzzy rule, the component is chosen for efficient design. When the criteria are not satisfied by the particular component, selection, crossover and mutation process is performed to choose the global optimal solution. By choosing the optimal component in GFOAM-NLADID Method, power consumption and design complexity level gets reduced. After designing the approximate multiplier, Non-Linear Anisotropic Diffusion Image Denoising process is carried out in GFOAM-NLADID Method to remove the noisy artifacts and to improve the image quality performance. This in turn helps to improve the peak signal-to-noise ratio performance by GFOAM-NLADID method. Experimental evaluation of GFOAM-NLADID Method is carried out with performance metrics such as power consumption, filtering time and peak signal-to-noise ratio compared to the state-of-the-art works.