HFirst: A Temporal Approach to Object Recognition

TL;DR

This paper presents a novel temporal, asynchronous neural model for object recognition that leverages precise timing information from event-based sensors, simplifying computation and achieving state-of-the-art accuracy in recognition tasks.

Contribution

It introduces a temporal winner-take-all approach using asynchronous AER sensors, shifting from traditional clocked systems to a timing-based paradigm for object recognition.

Findings

Achieved 97.5% accuracy on a four-class card pip recognition task.

Achieved 84.9% accuracy on a 36-class character recognition task.

Demonstrated the effectiveness of timing-based computation in visual recognition.

Abstract

This paper introduces a spiking hierarchical model for object recognition which utilizes the precise timing information inherently present in the output of biologically inspired asynchronous Address Event Representation (AER) vision sensors. The asynchronous nature of these systems frees computation and communication from the rigid predetermined timing enforced by system clocks in conventional systems. Freedom from rigid timing constraints opens the possibility of using true timing to our advantage in computation. We show not only how timing can be used in object recognition, but also how it can in fact simplify computation. Specifically, we rely on a simple temporal-winner-take-all rather than more computationally intensive synchronous operations typically used in biologically inspired neural networks for object recognition. This approach to visual computation represents a major paradigm shift from conventional clocked systems and can find application in other sensory modalities and computational tasks. We showcase effectiveness of the approach by achieving the highest reported accuracy to date (97.5\%$\pm$3.5\%) for a previously published four class card pip recognition task and an accuracy of 84.9\%$\pm$1.9\% for a new more difficult 36 class character recognition task.