Header

UZH-Logo

Maintenance Infos

Neuromorphic Approach Sensitivity Cell Modeling and FPGA Implementation


Liu, Hongjie; Rios-Navarro, Antonio; Moeys, Diederik Paul; Delbruck, Tobi; Linares-Barranco, Alejandro (2017). Neuromorphic Approach Sensitivity Cell Modeling and FPGA Implementation. In: International Conference on Artificial Neural Networks (ICANN) 2017, Alghero, Sardinia, Italy, 11 September 2017 - 14 September 2017, 179-187.

Abstract

Neuromorphic engineering takes inspiration from biology to solve engineering problems using the organizing principles of biological neural computation. This field has demonstrated success in sensor based applications (vision and audition) as well in cognition and actuators. This paper is focused on mimicking an interesting functionality of the retina that is computed by one type of Retinal Ganglion Cell (RGC). It is the early detection of approaching (expanding) dark objects. This paper presents the software and hardware logic FPGA implementation of this approach sensitivity cell. It can be used in later cognition layers as an attention mechanism. The input of this hardware modeled cell comes from an asynchronous spiking Dynamic Vision Sensor, which leads to an end-to-end event based processing system. The software model has been developed in Java, and computed with an average processing time per event of 370 ns on a NUC embedded computer. The output firing rate for an approaching object depends on the cell parameters that represent the needed number of input events to reach the firing threshold. For the hardware implementation on a Spartan6 FPGA, the processing time is reduced to 160 ns/event with the clock running at 50 MHz.

Abstract

Neuromorphic engineering takes inspiration from biology to solve engineering problems using the organizing principles of biological neural computation. This field has demonstrated success in sensor based applications (vision and audition) as well in cognition and actuators. This paper is focused on mimicking an interesting functionality of the retina that is computed by one type of Retinal Ganglion Cell (RGC). It is the early detection of approaching (expanding) dark objects. This paper presents the software and hardware logic FPGA implementation of this approach sensitivity cell. It can be used in later cognition layers as an attention mechanism. The input of this hardware modeled cell comes from an asynchronous spiking Dynamic Vision Sensor, which leads to an end-to-end event based processing system. The software model has been developed in Java, and computed with an average processing time per event of 370 ns on a NUC embedded computer. The output firing rate for an approaching object depends on the cell parameters that represent the needed number of input events to reach the firing threshold. For the hardware implementation on a Spartan6 FPGA, the processing time is reduced to 160 ns/event with the clock running at 50 MHz.

Statistics

Citations

Dimensions.ai Metrics

Altmetrics

Downloads

2 downloads since deposited on 23 Feb 2018
2 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Conference or Workshop Item (Paper), refereed, original work
Communities & Collections:07 Faculty of Science > Institute of Neuroinformatics
Dewey Decimal Classification:570 Life sciences; biology
Language:English
Event End Date:14 September 2017
Deposited On:23 Feb 2018 09:27
Last Modified:31 Jul 2018 05:11
Publisher:Artificial Neural Networks and Machine Learning – ICANN 2017
Series Name:International Conference on Artificial Neural Networks
Number of Pages:9
OA Status:Green
Publisher DOI:https://doi.org/10.1007/978-3-319-68600-4_22
Official URL:https://link.springer.com/chapter/10.1007%2F978-3-319-68600-4_22

Download

Download PDF  'Neuromorphic Approach Sensitivity Cell Modeling and FPGA Implementation'.
Preview
Content: Published Version
Filetype: PDF
Size: 26kB
View at publisher