Correction: Corrigendum: A spiking neural network model of 3D perception for event-based neuromorphic stereo vision systems

Osswald, Marc; Ieng, Sio-Hoï; Benosman, Ryad; Indiveri, Giacomo

doi:10.1038/srep44722

Cited by 4 publications

(22 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The SNN implemented on the DYNAP is adapted from the structure presented in Osswald et al (2017). It consists of three neuronal populations: the retina, the coincidence detectors, and the disparity detectors (see Figure 3).…”

Section: The Spiking Neural Network Modelmentioning

confidence: 99%

“…Moreover, disparity neurons receive feedforward lateral excitation from coincidence neurons tuned to the same disparity. This excitatory-inhibitory balance allows integrating the stimulus spatiotemporal features over time, thereby implementing the matching constraints of cooperative algorithms (Marr and Poggio, 1976;Mahowald, 1994b;Osswald et al, 2017). As a result, the SNN model can solve the stereo matching problem, with only TT represented in the disparity population.…”

Section: The Spiking Neural Network Modelmentioning

confidence: 99%

“…As shown in Osswald et al (2017), neurons in the emulated SNN model of cooperative stereo vision compute an approximation of the local covariance of the spatiotemporal visual information. As a result, neuronal and synaptic time constants are key parameters in the proposed architecture, and they were configured as follows.…”

Section: Network Calibrationmentioning

confidence: 99%

“…In this context, neuromorphic hardware has proven to be an effective substrate (Chicca et al, 2014;Indiveri et al, 2015). To date, the emerging field of event-based stereo vision has shown successful approaches that interface Spiking Neural Networks (SNNs) with neuromorphic event-based sensors, also referred to as "event cameras, " in order to build real-time event-based visual processing systems (Mahowald, 1994a;Osswald et al, 2017). Inspired by the retinal ganglion cells, the neuromorphic vision sensors broadcast information, independently for all the pixels, only in response to significant changes in illumination, which results in a low-power, low-latency, event-driven, and sparse input stream (Lichtsteiner et al, 2008;Posch et al, 2010;Berner et al, 2013).…”

Section: Introductionmentioning

confidence: 99%

“…Spiking neurons, in turn, can process signals using temporal information, and therefore, can take full advantage of an event-based input stream to solve the stereomatching problem. However, although several biologicallyinspired implementations of stereo vision (Mahowald, 1994b;Piatkowska et al, 2013Piatkowska et al, , 2017Dikov et al, 2017;Osswald et al, 2017;Kaiser et al, 2018) have extensively been explored, only a few solutions fully exploit the advantages of parallel computation, with an end-to-end neuromorphic architecture that can replace traditional Von Neumann architectures. In Dikov et al (2017), the first scalable architecture of the Marr and Poggio cooperative network (Marr and Poggio, 1976, 1977, 1979 is implemented on the SpiNNaker platform (Furber et al, 2014).…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A Spike-Based Neuromorphic Architecture of Stereo Vision

et al. 2020

Self Cite

View full text Add to dashboard Cite

The problem of finding stereo correspondences in binocular vision is solved effortlessly in nature and yet it is still a critical bottleneck for artificial machine vision systems. As temporal information is a crucial feature in this process, the advent of event-based vision sensors and dedicated event-based processors promises to offer an effective approach to solving the stereo matching problem. Indeed, event-based neuromorphic hardware provides an optimal substrate for fast, asynchronous computation, that can make explicit use of precise temporal coincidences. However, although several biologically-inspired solutions have already been proposed, the performance benefits of combining event-based sensing with asynchronous and parallel computation are yet to be explored. Here we present a hardware spike-based stereo-vision system that leverages the advantages of brain-inspired neuromorphic computing by interfacing two event-based vision sensors to an event-based mixed-signal analog/digital neuromorphic processor. We describe a prototype interface designed to enable the emulation of a stereo-vision system on neuromorphic hardware and we quantify the stereo matching performance with two datasets. Our results provide a path toward the realization of low-latency, end-to-end event-based, neuromorphic architectures for stereo vision.

show abstract

Section: The Spiking Neural Network Modelmentioning

confidence: 99%

Section: The Spiking Neural Network Modelmentioning

confidence: 99%

Section: Network Calibrationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A Spike-Based Neuromorphic Architecture of Stereo Vision

et al. 2020

Self Cite

View full text Add to dashboard Cite

show abstract

Microsaccades for Neuromorphic Stereo Vision

Kaiser

Weinland

Keller

et al. 2018

Artificial Neural Networks and Machine Learning – ICANN 2018

View full text Add to dashboard Cite

Instantaneous Stereo Depth Estimation of Real-World Stimuli with a Neuromorphic Stereo-Vision Setup

Risi

Calabrese

Indiveri

2021

2021 IEEE International Symposium on Circuits and Systems (ISCAS)

Self Cite

View full text Add to dashboard Cite

The stereo-matching problem, i.e., matching corresponding features in two different views to reconstruct depth, is efficiently solved in biology. Yet, it remains the computational bottleneck for classical machine vision approaches. By exploiting the properties of event cameras, recently proposed Spiking Neural Network (SNN) architectures for stereo vision have the potential of simplifying the stereo-matching problem. Several solutions that combine event cameras with spike-based neuromorphic processors already exist. However, they are either simulated on digital hardware or tested on simplified stimuli. In this work, we use the Dynamic Vision Sensor 3D Human Pose Dataset (DHP19) to validate a brain-inspired event-based stereo-matching architecture implemented on a mixed-signal neuromorphic processor with real-world data. Our experiments show that this SNN architecture, composed of coincidence detectors and disparity sensitive neurons, is able to provide a coarse estimate of the input disparity instantaneously, thereby detecting the presence of a stimulus moving in depth in real-time.Index Terms-event-based, 3D dataset, mixed-signal hardware, analog circuits, spiking neural networks, disparity.

show abstract

Correction: Corrigendum: A spiking neural network model of 3D perception for event-based neuromorphic stereo vision systems

Cited by 4 publications

References 1 publication

A Spike-Based Neuromorphic Architecture of Stereo Vision

A Spike-Based Neuromorphic Architecture of Stereo Vision

Microsaccades for Neuromorphic Stereo Vision

Instantaneous Stereo Depth Estimation of Real-World Stimuli with a Neuromorphic Stereo-Vision Setup

Contact Info

Product

Resources

About