Memristor Technologies for Dynamic Vision Sensors: A Critical Assessment and Research Roadmap

TL;DR

This paper critically assesses memristor-based dynamic vision sensors, highlighting their potential to revolutionize edge-AI by enabling low-power, in-memory analog computation, and outlines a research roadmap for practical implementation.

Contribution

It provides a comprehensive review of memristor technologies for vision sensors, introduces a three-paradigm architectural taxonomy, and identifies key challenges and benchmarks for future research.

Findings

Existing hardware is at TRL 2-5, mostly in projection.

Half of surveyed applications rely on projections, not real systems.

End-to-end DVS-memristor systems are the main open challenge.

Abstract

Edge-AI deployment is bottlenecked by data-movement energy; pairing event-driven vision sensors with in-memory analog compute could lift that ceiling by orders of magnitude. Both technologies are individually mature; the framework distinguishing fabricated demonstrations from projected systems is missing. Of six application domains surveyed (robotics, autonomous vehicles, AR/VR, surveillance, medical imaging, IoT), half rest entirely on projection, and existing hardware sits at Technology Readiness Levels 2-5. This evidence-graded review applies a three-paradigm architectural taxonomy and benchmarks the gap against current digital neuromorphic alternatives. It identifies an end-to-end integrated DVS-memristor system as the field's open challenge, with testable accuracy and power targets.