Display options
Cite
Cohen GK, Orchard G, Leng SH, et al. Skimming Digits: Neuromorphic Classification of Spike-Encoded Images. Front Neurosci. 2016;10:184doi: 10.3389/fnins.2016.00184.
Cohen, G. K., Orchard, G., Leng, S. H., Tapson, J., Benosman, R. B., & van Schaik, A. (2016). Skimming Digits: Neuromorphic Classification of Spike-Encoded Images. Frontiers in neuroscience, 10184. https://doi.org/10.3389/fnins.2016.00184
Cohen, Gregory K, et al. "Skimming Digits: Neuromorphic Classification of Spike-Encoded Images." Frontiers in neuroscience vol. 10 (2016): 184. doi: https://doi.org/10.3389/fnins.2016.00184
Cohen GK, Orchard G, Leng SH, Tapson J, Benosman RB, van Schaik A. Skimming Digits: Neuromorphic Classification of Spike-Encoded Images. Front Neurosci. 2016 Apr 28;10:184. doi: 10.3389/fnins.2016.00184. eCollection 2016. PMID: 27199646; PMCID: PMC4848313.
Copy Download .nbib Format: NLM
Share it on

Front Neurosci. 2016 Apr 28;10:184. doi: 10.3389/fnins.2016.00184. eCollection 2016.

Skimming Digits: Neuromorphic Classification of Spike-Encoded Images.

Frontiers in neuroscience

Gregory K Cohen, Garrick Orchard, Sio-Hoi Leng, Jonathan Tapson, Ryad B Benosman, André van Schaik

Affiliations

  1. Biomedical Engineering and Neuroscience, The MARCS Institute, Western Sydney UniversitySydney, NSW, Australia; Natural Vision and Computation Team, Vision Institute, University Pierre and Marie Curie-Centre National de la Recherche ScientifiqueParis, France.
  2. Temasek Labs (TLAB), National University of SingaporeSingapore, Singapore; Neuromorphic Engineering and Robotics, Singapore Institute for Neurotechnology (SINAPSE), National University of SingaporeSingapore, Singapore.
  3. Natural Vision and Computation Team, Vision Institute, University Pierre and Marie Curie-Centre National de la Recherche Scientifique Paris, France.
  4. Biomedical Engineering and Neuroscience, The MARCS Institute, Western Sydney University Sydney, NSW, Australia.

PMID: 27199646 PMCID: PMC4848313 DOI: 10.3389/fnins.2016.00184

Abstract

The growing demands placed upon the field of computer vision have renewed the focus on alternative visual scene representations and processing paradigms. Silicon retinea provide an alternative means of imaging the visual environment, and produce frame-free spatio-temporal data. This paper presents an investigation into event-based digit classification using N-MNIST, a neuromorphic dataset created with a silicon retina, and the Synaptic Kernel Inverse Method (SKIM), a learning method based on principles of dendritic computation. As this work represents the first large-scale and multi-class classification task performed using the SKIM network, it explores different training patterns and output determination methods necessary to extend the original SKIM method to support multi-class problems. Making use of SKIM networks applied to real-world datasets, implementing the largest hidden layer sizes and simultaneously training the largest number of output neurons, the classification system achieved a best-case accuracy of 92.87% for a network containing 10,000 hidden layer neurons. These results represent the highest accuracies achieved against the dataset to date and serve to validate the application of the SKIM method to event-based visual classification tasks. Additionally, the study found that using a square pulse as the supervisory training signal produced the highest accuracy for most output determination methods, but the results also demonstrate that an exponential pattern is better suited to hardware implementations as it makes use of the simplest output determination method based on the maximum value.

Keywords: N-MNIST; OPIUM; SKIM; multi-class; object classification

References

  1. Front Neurosci. 2015 Nov 16;9:437 - PubMed
  2. IEEE Trans Pattern Anal Mach Intell. 2015 Oct;37(10):2028-40 - PubMed
  3. IEEE Trans Pattern Anal Mach Intell. 2013 Nov;35(11):2706-19 - PubMed
  4. Front Neurosci. 2013 Aug 30;7:153 - PubMed
  5. Neural Comput. 2015 Mar;27(3):561-93 - PubMed
  6. Front Neurosci. 2013 Oct 08;7:178 - PubMed
  7. Science. 2014 Aug 8;345(6197):668-73 - PubMed

Publication Types