Pose-invariant object recognition for event-based vision with slow-ELM

Rohan Ghosh; Tang Siyi; Mahdi Rasouli; Nitish V. Thakor; Sunil L. Kukreja

doi:10.1007/978-3-319-44781-0_54

Pose-invariant object recognition for event-based vision with slow-ELM

Rohan Ghosh, Tang Siyi, Mahdi Rasouli, Nitish V. Thakor, Sunil L. Kukreja

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Neuromorphic image sensors produce activity-driven spiking output at every pixel. These low-power consuming imagers which encode visual change information in the form of spikes help reduce computational overhead and realize complex real-time systems; object recognition and pose-estimation to name a few. However, there exists a lack of algorithms in event-based vision aimed towards capturing invariance to transformations. In this work, we propose a methodology for recognizing objects invariant to their pose with the Dynamic Vision Sensor (DVS). A novel slow-ELM architecture is proposed which combines the effectiveness of Extreme Learning Machines and Slow Feature Analysis. The system, tested on an Intel Core i5-4590 CPU, can perform 10, 000 classifications per second and achieves 1% classification error for 8 objects with views accumulated over 90° of 2D pose.

Original language	English (US)
Title of host publication	Artificial Neural Networks and Machine Learning - 25th International Conference on Artificial Neural Networks, ICANN 2016, Proceedings
Editors	Alessandro E.P. Villa, Paolo Masulli, Antonio Javier Pons Rivero
Publisher	Springer Verlag
Pages	455-462
Number of pages	8
ISBN (Print)	9783319447803
DOIs	https://doi.org/10.1007/978-3-319-44781-0_54
State	Published - 2016
Externally published	Yes
Event	25th International Conference on Artificial Neural Networks and Machine Learning, ICANN 2016 - Barcelona, Spain Duration: Sep 6 2016 → Sep 9 2016

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9887 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	25th International Conference on Artificial Neural Networks and Machine Learning, ICANN 2016
Country/Territory	Spain
City	Barcelona
Period	9/6/16 → 9/9/16

Keywords

Extreme learning machines
Neuromorphic vision
Object recognition
Slow feature analysis

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-319-44781-0_54

Cite this

Ghosh, R., Siyi, T., Rasouli, M., Thakor, N. V., & Kukreja, S. L. (2016). Pose-invariant object recognition for event-based vision with slow-ELM. In A. E. P. Villa, P. Masulli, & A. J. P. Rivero (Eds.), Artificial Neural Networks and Machine Learning - 25th International Conference on Artificial Neural Networks, ICANN 2016, Proceedings (pp. 455-462). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9887 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-44781-0_54

Pose-invariant object recognition for event-based vision with slow-ELM. / Ghosh, Rohan; Siyi, Tang; Rasouli, Mahdi et al.
Artificial Neural Networks and Machine Learning - 25th International Conference on Artificial Neural Networks, ICANN 2016, Proceedings. ed. / Alessandro E.P. Villa; Paolo Masulli; Antonio Javier Pons Rivero. Springer Verlag, 2016. p. 455-462 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9887 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Ghosh, R, Siyi, T, Rasouli, M, Thakor, NV & Kukreja, SL 2016, Pose-invariant object recognition for event-based vision with slow-ELM. in AEP Villa, P Masulli & AJP Rivero (eds), Artificial Neural Networks and Machine Learning - 25th International Conference on Artificial Neural Networks, ICANN 2016, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9887 LNCS, Springer Verlag, pp. 455-462, 25th International Conference on Artificial Neural Networks and Machine Learning, ICANN 2016, Barcelona, Spain, 9/6/16. https://doi.org/10.1007/978-3-319-44781-0_54

Ghosh R, Siyi T, Rasouli M, Thakor NV, Kukreja SL. Pose-invariant object recognition for event-based vision with slow-ELM. In Villa AEP, Masulli P, Rivero AJP, editors, Artificial Neural Networks and Machine Learning - 25th International Conference on Artificial Neural Networks, ICANN 2016, Proceedings. Springer Verlag. 2016. p. 455-462. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-44781-0_54

Ghosh, Rohan ; Siyi, Tang ; Rasouli, Mahdi et al. / Pose-invariant object recognition for event-based vision with slow-ELM. Artificial Neural Networks and Machine Learning - 25th International Conference on Artificial Neural Networks, ICANN 2016, Proceedings. editor / Alessandro E.P. Villa ; Paolo Masulli ; Antonio Javier Pons Rivero. Springer Verlag, 2016. pp. 455-462 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{352c09e5f802417ab04b5f84264a6770,

title = "Pose-invariant object recognition for event-based vision with slow-ELM",

abstract = "Neuromorphic image sensors produce activity-driven spiking output at every pixel. These low-power consuming imagers which encode visual change information in the form of spikes help reduce computational overhead and realize complex real-time systems; object recognition and pose-estimation to name a few. However, there exists a lack of algorithms in event-based vision aimed towards capturing invariance to transformations. In this work, we propose a methodology for recognizing objects invariant to their pose with the Dynamic Vision Sensor (DVS). A novel slow-ELM architecture is proposed which combines the effectiveness of Extreme Learning Machines and Slow Feature Analysis. The system, tested on an Intel Core i5-4590 CPU, can perform 10, 000 classifications per second and achieves 1% classification error for 8 objects with views accumulated over 90° of 2D pose.",

keywords = "Extreme learning machines, Neuromorphic vision, Object recognition, Slow feature analysis",

author = "Rohan Ghosh and Tang Siyi and Mahdi Rasouli and Thakor, {Nitish V.} and Kukreja, {Sunil L.}",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2016.; 25th International Conference on Artificial Neural Networks and Machine Learning, ICANN 2016 ; Conference date: 06-09-2016 Through 09-09-2016",

year = "2016",

doi = "10.1007/978-3-319-44781-0_54",

language = "English (US)",

isbn = "9783319447803",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "455--462",

editor = "Villa, {Alessandro E.P.} and Paolo Masulli and Rivero, {Antonio Javier Pons}",

booktitle = "Artificial Neural Networks and Machine Learning - 25th International Conference on Artificial Neural Networks, ICANN 2016, Proceedings",

}

TY - GEN

T1 - Pose-invariant object recognition for event-based vision with slow-ELM

AU - Ghosh, Rohan

AU - Siyi, Tang

AU - Rasouli, Mahdi

AU - Thakor, Nitish V.

AU - Kukreja, Sunil L.

N1 - Publisher Copyright: © Springer International Publishing Switzerland 2016.

PY - 2016

Y1 - 2016

N2 - Neuromorphic image sensors produce activity-driven spiking output at every pixel. These low-power consuming imagers which encode visual change information in the form of spikes help reduce computational overhead and realize complex real-time systems; object recognition and pose-estimation to name a few. However, there exists a lack of algorithms in event-based vision aimed towards capturing invariance to transformations. In this work, we propose a methodology for recognizing objects invariant to their pose with the Dynamic Vision Sensor (DVS). A novel slow-ELM architecture is proposed which combines the effectiveness of Extreme Learning Machines and Slow Feature Analysis. The system, tested on an Intel Core i5-4590 CPU, can perform 10, 000 classifications per second and achieves 1% classification error for 8 objects with views accumulated over 90° of 2D pose.

AB - Neuromorphic image sensors produce activity-driven spiking output at every pixel. These low-power consuming imagers which encode visual change information in the form of spikes help reduce computational overhead and realize complex real-time systems; object recognition and pose-estimation to name a few. However, there exists a lack of algorithms in event-based vision aimed towards capturing invariance to transformations. In this work, we propose a methodology for recognizing objects invariant to their pose with the Dynamic Vision Sensor (DVS). A novel slow-ELM architecture is proposed which combines the effectiveness of Extreme Learning Machines and Slow Feature Analysis. The system, tested on an Intel Core i5-4590 CPU, can perform 10, 000 classifications per second and achieves 1% classification error for 8 objects with views accumulated over 90° of 2D pose.

KW - Extreme learning machines

KW - Neuromorphic vision

KW - Object recognition

KW - Slow feature analysis

UR - http://www.scopus.com/inward/record.url?scp=84988432083&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84988432083&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-44781-0_54

DO - 10.1007/978-3-319-44781-0_54

M3 - Conference contribution

AN - SCOPUS:84988432083

SN - 9783319447803

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 455

EP - 462

BT - Artificial Neural Networks and Machine Learning - 25th International Conference on Artificial Neural Networks, ICANN 2016, Proceedings

A2 - Villa, Alessandro E.P.

A2 - Masulli, Paolo

A2 - Rivero, Antonio Javier Pons

PB - Springer Verlag

T2 - 25th International Conference on Artificial Neural Networks and Machine Learning, ICANN 2016

Y2 - 6 September 2016 through 9 September 2016

ER -

Pose-invariant object recognition for event-based vision with slow-ELM

Abstract

Publication series

Other

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this