350 rub
Journal Neurocomputers №2 for 2017 г.
Article in number:
Methods of preparing the data for processing by spiking neuron networks
Authors:
V.I. Terekhov - Ph.D. (Eng.), Associate Professor, Department «Information Processing and Control Systems», Bauman Moscow State Technical University E-mail: terekchow@bmstu.ru R.V. Zhukov - Post-graduate Student, Department «Information Processing and Control Systems», Bauman Moscow State Technical University E-mail: zhukov@student.bmstu.ru
Abstract:
This paper presents a data preparation method for its processing by a receptor layer using an artificial neuron network. The receptor layer transforms the input data into an impulse series and, therefore, it is an essential part of the impulse neural network. It is necessary to create a group of neurons for building the receptor layer using one of the known models of the impulse neuron. The paper proposes to use the LIF-model, since it is the simplest. A dependence of the output frequency of the output tension defines for every neuron of the receptor layer. The dependence is proportional to the signal value which is transferred to the entry of the neuron. It is shown that such an approach can be applied to the input data of different types. These inputs can be: a numerical vector, images, video stream and data from various sensors. The propose methodic for preparation and converting the original data brings data into a sequence of numeric values which are necessary to sort out or find out parameters for impulse neurons of the receptor layer for required input values diapasons. The described methodic of building receptor layers of neural networks can significantly simplify impulse neuron application for many practical tasks.
Pages: 31-36
References

 

  1. Trhan P. The application of spiking neural networks in autonomous robot control //Computing and Informatics. 2012. T. 29. № 5. S. 823-847.
  2. Martin H., Conradt J. Spiking neural networks for vision tasks. 2015.
  3. Huang L. et al. Circle detection using a spiking neural network // Image and Signal Processing (CISP), 2013 6th International Congress on // IEEE. 2013. T. 3. S. 1442-1446.
  4. Goodman D., Brette R. Brian: a simulator for spiking neural networks in Python. 2008.
  5. Orhan E. The Leaky Integrate-and-Fire Neuron Model. 2012.
  6. PNG images and cliparts for web design URL: http://pngimg.com/ (data obrashhenija: 20.12.2016).
  7. The Open Video Project URL: http://memory.loc.gov/mbrs/varsmp/0763.mpg (data obrashhenija: 25.12.2016).
  8. The Open Video Project URL: http://memory.loc.gov/mbrs/awal/0868.mpg (data obrashhenija: 25.12.2016).
  9. Anguita D. et al. A Public Domain Dataset for Human Activity Recognition using Smartphones // ESANN. 2013.