The design of image low-pass filter ASIC architecture based on stochastic computing
Xueru Zhao, Guoqiang He, Kejun Ge · 2024
The image low-pass filter is an effective solution for denoising and improving image quality. The stochastic computing is a novel method for number representation and calculation in stochastic domain. Aiming at the problem of large computing resources and low operation speed in image low-pass filter processing, in this paper, a parallel ASIC architecture of image low-pass filter based on stochastic computing proposed, which effectively improves processing speed and reduces resource consumption. When the length of the probability domain is 128bits, compared with the traditional architecture, the computational gate resources are reduced by about 30% and the execution time is reduced by about 75% in the case of comparable processing performance.