An Energy-Efficient VCO-Based Matrix Multiplier Block to Support On-Chip Image Analysis

TL;DR

This paper introduces an energy-efficient, time-domain analog MAC architecture for on-chip image analysis, achieving 400 times lower power consumption than digital methods, suitable for scaled CMOS technologies.

Contribution

It presents a novel time-domain signal processing architecture for MAC operations that significantly reduces energy consumption in on-chip image analysis.

Findings

Achieves 7-bit accuracy in MAC operations.

Consumes 400 times less energy than digital implementations.

Suitable for scaled CMOS technologies.

Abstract

Images typically are represented as uniformly sampled data in the form of matrix of pixels/voxels. Therefore, matrix multiply-and-accumulate (MAC) forms the core of most state-of-the-art image analysis algorithms. While digital implementation of MAC has generally been the preferred approach, high power consumption is an impediment to adopting it for medical image analysis. In this work, we present a time-domain signal processing architecture which performs MAC operations with 7bit accuracy while consuming 400X lower energy than digital implementation. The proposed architecture performs analog computation using mostly digital circuits and is suitable for scaled CMOS technologies. The proposed time-domain MAC architecture is expected to play a central role in empowering the advancement of various on-chip image analysis operations.