Two Efficient Approximate Unsigned Multipliers by Developing New Configuration for Approximate 4:2 Compressors

Approximate computing is a promising approach for reducing power consumption and design complexity in applications that accuracy not crucial factor. multipliers are commonly used error-tolerant applications. This paper presents three approximate 4:2 compressors two multiplier designs, aiming at the area consumption, while maintaining acceptable accuracy. The seeks to develop align positive negative approximations input patterns have same probability. Additionally, proposed utilized construct different columns of partial products based on probabilities adjacent columns. synthesized using 28nm technology. Compared exact multiplier, first improves $\times $ delay by 91% 86%, respectively, second parameters 90% 84%, respectively. performance methods was assessed compared with existing image multiplication, sharpening, smoothing edge detection. Also, hardware implementation neural network investigated, simulation results indicate appropriate these