Near-Memory Architecture for Threshold-Ordinal Surface-Based Corner Detection of Event Cameras

TL;DR

This paper introduces a near-memory hardware architecture for efficient threshold-ordinal surface updates in event camera corner detection, significantly reducing latency and energy consumption on resource-limited edge devices.

Contribution

It proposes a novel near-memory architecture with optimized SRAM and power management techniques for fast, low-power corner detection in event-based cameras.

Findings

Reduces latency by up to 24.7x and energy by 1.2x at 1.2 V

Achieves robust circuit operation with minimal bit errors above 0.62 V

Maintains high corner detection accuracy with minor AUC reduction

Abstract

Event-based Cameras (EBCs) are widely utilized in surveillance and autonomous driving applications due to their high speed and low power consumption. Corners are essential low-level features in event-driven computer vision, and novel algorithms utilizing event-based representations, such as Threshold-Ordinal Surface (TOS), have been developed for corner detection. However, the implementation of these algorithms on resource-constrained edge devices is hindered by significant latency, undermining the advantages of EBCs. To address this challenge, a near-memory architecture for efficient TOS updates (NM-TOS) is proposed. This architecture employs a read-write decoupled 8T SRAM cell and optimizes patch update speed through pipelining. Hardware-software co-optimized peripheral circuits and dynamic voltage and frequency scaling (DVFS) enable power and latency reductions. Compared to traditional digital implementations, our architecture reduces latency/energy by 24.7x/1.2x at Vdd = 1.2 V or 1.93x/6.6x at Vdd = 0.6 V based on 65nm CMOS process. Monte Carlo simulations confirm robust circuit operation, demonstrating zero bit error rate at operating voltages above 0.62 V, with only 0.2% at 0.61 V and 2.5% at 0.6 V. Corner detection evaluation using precision-recall area under curve (AUC) metrics reveals minor AUC reductions of 0.027 and 0.015 at 0.6 V for two popular EBC datasets.