Review of Approximate Computing in Image Processing Applications
Surbhi Bharti, Ashwni Kumar, Pankaj Gupta
Abstract
Quality of real time multimedia images mostly depends on the input, i.e. size of camera sensor, pixel density, quality of lens, etc. therefore the output sometimes contains unsolvable images issues and requires image processing. Approximate computing (AC) has emerged as innovative method for image processing along with the advantages of lower power and high accuracy. In this paper, a review of AC application in image processing is done and a bit truncated Pseudo-8T Static Random Access Memory (BT-PSRAM) is proposed. It is implemented on 32nm CMOS technology node for compressed multimedia applications, and gives encouraging result with low power and energy consumption. Size of Pseudo-8T SRAM has been reduced using 2-bit truncation technique.