A 64x64 aer logarithmic temporal derivative silicon retina

Lichtsteiner, P.; Delbrück, Tobi

doi:10.1109/rme.2005.1542972

Cited by 61 publications

(64 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This method has two main advantages when used in SNN: (1) fast learning (as the order of the first incoming spikes is often sufficient information for recognising a pattern and for a fast decision making and only one pass propagation of the input pattern may be sufficient for the model to learn it); (2) asynchronous, data-driven processing. As a consequence, RO learning is most appropriate for AER input data streams as the address-events are coneyed into the SNN 'one by one', in the order of their happening (Lichtsteiner and Delbruck, 2005;Delbruck, 2007). Thorpe, S. and J. Gautrais (1998) utilised RO learning to achieve fast, one-pass learning of static patterns (images).…”

Section: Evolving Spiking Neural Network (Esnn)mentioning

confidence: 99%

“…: address event representation (AER) devices, such as the artificial retina (Lichtsteiner and Delbruck, 2005;Delbruck, 2007) and artificial cochlea (van Shaik and Liu, 2005), the available wireless EEG equipment (e.g. Emotive) and with the advanced SNN hardware technologies (Indiveri et al, -2011, new opportunities have been created, but this still requires efficient and suitable methods.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Dynamic evolving spiking neural networks for on-line spatio- and spectro-temporal pattern recognition

et al. 2013

View full text Add to dashboard Cite

On-line learning and recognition of spatio-and spectro-temporal data (SSTD) is a very challenging task and an important one for the future development of autonomous machine learning systems with broad applications. Models based on spiking neural networks (SNN) have already proved their potential in capturing spatial and temporal data. One class of them, the evolving SNN (eSNN), uses a one-pass rank-order learning mechanism and a strategy to evolve a new spiking neuron and new connections to learn new patterns from incoming data. So far these networks have been mainly used for fast image and speech frame-based recognition. Alternative spike-time learning methods, such as SpikeTiming Dependent Plasticity (STDP) and its variant Spike Driven Synaptic Plasticity (SDSP), can also be used to learn spatio-temporal representations, but they usually require many iterations in an unsupervised or semi-supervised mode of learning. This paper introduces a new class of eSNN, dynamic eSNN, that utilise both rank-order learning and dynamic synapses to learn SSTD in a fast, on-line mode. The paper also introduces a new model called deSNN, that utilises rank-order learning and SDSP spike-time learning in unsupervised, supervised, or semi-supervised modes. The SDSP learning is used to evolve dynamically the network changing connection weights that capture spatio-temporal spike data clusters both during training and during recall. The new deSNN model is first illustrated on simple examples and then applied on two case study applications: (1) moving object recognition using address-event representation (AER) with data collected using a silicon retina device; (2) EEG SSTD recognition for brain-computer interfaces. The deSNN models resulted in a superior performance in terms of accuracy and speed when compared with other SNN models that use either rank-order or STDP learning. The reason is that the deSNN makes use of both the information contained in the order of the first input spikes (which information is explicitly present in input data streams and would be crucial to consider in some tasks) and of the information contained in the timing of the following spikes that is learned by the dynamic synapses as a whole spatio-temporal pattern. The paper is published in Neural Networks, Elsevier, e- AbstractOn-line learning and recognition of spatio-and spectro-temporal data (SSTD) is a very challenging task and an important one for the future development of autonomous machine learning systems with broad applications. Models based on spiking neural networks (SNN) have already proved their potential in capturing spatial and temporal data. One class of them, the evolving SNN (eSNN), uses a one-pass rank-order learning mechanism and a strategy to evolve a new spiking neuron and new connections to learn new patterns from incoming data. So far these networks have been mainly used for fast image and speech frame-based recognition. Alternative spike-time learning methods, such as Spike-Timing Dependent Plasticity (STDP) and its variant S...

show abstract

Section: Evolving Spiking Neural Network (Esnn)mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Dynamic evolving spiking neural networks for on-line spatio- and spectro-temporal pattern recognition

et al. 2013

View full text Add to dashboard Cite

show abstract

“…The chip has a one-dimensional array of 24 motion pixels that implement a time-to-travel algorithm. Each motion pixel contains a photodiode and an analog photoreceptor circuit [14] that convert a temporal change of light intensity into a transient voltage signal. The output signal is further amplified and high-pass filtered by a circuit descibed in [12].…”

Section: Motion Detection Chipmentioning

confidence: 99%

“…Various schemes have been adopted on chip to reduce the dependence of the computed motion on background intensity and contrast. The computed motion is invariant to changes in background intensity over at least three decades due to the inclusion of the adaptive photoreceptor circuits in [14]. The computed motion is also invariant to image contrast, C down to values of 2.5% due to the contrast edge detection circuits.…”

Section: Introductionmentioning

confidence: 95%

Steering with an aVLSI motion detection chip

Moeckel

Jaeggi

Liu

2008

2008 IEEE International Symposium on Circuits and Systems (ISCAS)

View full text Add to dashboard Cite

“…The core components are a temporal-derivative retina chip that encodes temporal contrast changes in spike output [1], and a multi-neuron chip which contains the neurons of the classifier [3]. Spikes are transmitted between the two chips using an asynchronous, event-based real-time communication protocol (address-event representation, AER), implemented by a framework of digital AER components [2].…”

Section: Introductionmentioning

confidence: 99%

A Spike-Based Saccadic Recognition System

Oster¹,

Lichtsteiner²,

Delbrück³

et al. 2007

2007 IEEE International Symposium on Circuits and Systems (ISCAS)

View full text Add to dashboard Cite

Abstract-We present a spike-based saccadic recognition system that uses a temporal-derivative silicon retina on a pan-tilt unit and an aVLSI multi-neuron classifier with a time-to-firstspike output coding. By using the spike information during the last 150 ms of a saccadic movement, we generate a reliable, sparse stimulus representation of image patches. We describe a novel classification scheme where the retinal spikes during this time influence the time-to-first spike of classifier neurons which receive the same constant input current. The preferred pattern of the neuron is stored in the synaptic connectivity between the retina and the classifier neuron. We demonstrate the robustness and real-time performance of this recognition scheme on a saccadic system which uses analog VLSI components.

show abstract

A 64x64 aer logarithmic temporal derivative silicon retina

Cited by 61 publications

References 4 publications

Dynamic evolving spiking neural networks for on-line spatio- and spectro-temporal pattern recognition

Dynamic evolving spiking neural networks for on-line spatio- and spectro-temporal pattern recognition

Steering with an aVLSI motion detection chip

A Spike-Based Saccadic Recognition System

Contact Info

Product

Resources

About