Deep Learning for Semantic Segmentation on Minimal Hardware

Sander G. van Dijk; Marcus M. Scheunemann

doi:10.1007/978-3-030-27544-0_29

Deep Learning for Semantic Segmentation on Minimal Hardware

Sander G. van Dijk, Marcus M. Scheunemann

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

3 Citations (Scopus)

16 Downloads (Pure)

Abstract

Deep learning has revolutionised many fields, but it is still challenging to transfer its success to small mobile robots with minimal hardware. Specifically, some work has been done to this effect in the RoboCup humanoid football domain, but results that are performant and efficient and still generally applicable outside of this domain are lacking. We propose an approach conceptually different from those taken previously. It is based on semantic segmentation and does achieve these desired properties. In detail, it is being able to process full VGA images in real-time on a low-power mobile processor. It can further handle multiple image dimensions without retraining, it does not require specific domain knowledge to achieve a high frame rate and it is applicable on a minimal mobile hardware.

Original language	English
Title of host publication	RoboCup 2018
Subtitle of host publication	Robot World Cup XXII
Editors	Dirk Holz, Katie Genter, Maarouf Saad, Oskar von Stryk
Publisher	Springer Nature
Pages	349-361
Number of pages	13
ISBN (Electronic)	9783030275440
ISBN (Print)	9783030275433
DOIs	https://doi.org/10.1007/978-3-030-27544-0_29
Publication status	Published - 4 Aug 2019
Event	RoboCup 2018 Symposium - Palais des congrès, Montreal, Canada Duration: 22 Jun 2018 → 22 Jun 2018 http://www.robocup2018.org/?page=symposium&lang=en

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11374 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	RoboCup 2018 Symposium
Country/Territory	Canada
City	Montreal
Period	22/06/18 → 22/06/18
Internet address	http://www.robocup2018.org/?page=symposium&lang=en

Keywords

Robotics
Machine Learning
Deep Learning
Computer Vision
Semantic Segmentation
Minimal Hardware
Mobile Robotics
Deep learning
Computer vision
Minimal hardware
Semantic segmentation
Mobile robotics

Access to Document

10.1007/978-3-030-27544-0_29

2018-Deep-Learning-for-Semantic-Segmentation-on-Minimal-Hardware--1Accepted author manuscript, 3.21 MBLicence: Other

Cite this

Dijk, S. G. V., & Scheunemann, M. M. (2019). Deep Learning for Semantic Segmentation on Minimal Hardware. In D. Holz, K. Genter, M. Saad, & O. von Stryk (Eds.), RoboCup 2018: Robot World Cup XXII (pp. 349-361). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11374 LNAI). Springer Nature. https://doi.org/10.1007/978-3-030-27544-0_29

Dijk, Sander G. van ; Scheunemann, Marcus M. / Deep Learning for Semantic Segmentation on Minimal Hardware. RoboCup 2018: Robot World Cup XXII. editor / Dirk Holz ; Katie Genter ; Maarouf Saad ; Oskar von Stryk. Springer Nature, 2019. pp. 349-361 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{b7197f40dc8e4d9cb4cac3d5993403cc,

title = "Deep Learning for Semantic Segmentation on Minimal Hardware",

abstract = "Deep learning has revolutionised many fields, but it is still challenging to transfer its success to small mobile robots with minimal hardware. Specifically, some work has been done to this effect in the RoboCup humanoid football domain, but results that are performant and efficient and still generally applicable outside of this domain are lacking. We propose an approach conceptually different from those taken previously. It is based on semantic segmentation and does achieve these desired properties. In detail, it is being able to process full VGA images in real-time on a low-power mobile processor. It can further handle multiple image dimensions without retraining, it does not require specific domain knowledge to achieve a high frame rate and it is applicable on a minimal mobile hardware.",

keywords = "Robotics, Machine Learning, Deep Learning, Computer Vision, Semantic Segmentation, Minimal Hardware, Mobile Robotics, Deep learning, Computer vision, Minimal hardware, Semantic segmentation, Mobile robotics",

author = "Dijk, {Sander G. van} and Scheunemann, {Marcus M.}",

note = "{\textcopyright} Springer Nature Switzerland AG 2019; RoboCup 2018 Symposium ; Conference date: 22-06-2018 Through 22-06-2018",

year = "2019",

month = aug,

day = "4",

doi = "10.1007/978-3-030-27544-0_29",

language = "English",

isbn = "9783030275433",

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

publisher = "Springer Nature",

pages = "349--361",

editor = "Dirk Holz and Katie Genter and Maarouf Saad and {von Stryk}, Oskar",

booktitle = "RoboCup 2018",

address = "Netherlands",

url = "http://www.robocup2018.org/?page=symposium&lang=en",

}

Dijk, SGV & Scheunemann, MM 2019, Deep Learning for Semantic Segmentation on Minimal Hardware. in D Holz, K Genter, M Saad & O von Stryk (eds), RoboCup 2018: Robot World Cup XXII. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11374 LNAI, Springer Nature, pp. 349-361, RoboCup 2018 Symposium, Montreal, Quebec, Canada, 22/06/18. https://doi.org/10.1007/978-3-030-27544-0_29

Deep Learning for Semantic Segmentation on Minimal Hardware. / Dijk, Sander G. van; Scheunemann, Marcus M.
RoboCup 2018: Robot World Cup XXII. ed. / Dirk Holz; Katie Genter; Maarouf Saad; Oskar von Stryk. Springer Nature, 2019. p. 349-361 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11374 LNAI).

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

TY - GEN

T1 - Deep Learning for Semantic Segmentation on Minimal Hardware

AU - Dijk, Sander G. van

AU - Scheunemann, Marcus M.

N1 - © Springer Nature Switzerland AG 2019

PY - 2019/8/4

Y1 - 2019/8/4

N2 - Deep learning has revolutionised many fields, but it is still challenging to transfer its success to small mobile robots with minimal hardware. Specifically, some work has been done to this effect in the RoboCup humanoid football domain, but results that are performant and efficient and still generally applicable outside of this domain are lacking. We propose an approach conceptually different from those taken previously. It is based on semantic segmentation and does achieve these desired properties. In detail, it is being able to process full VGA images in real-time on a low-power mobile processor. It can further handle multiple image dimensions without retraining, it does not require specific domain knowledge to achieve a high frame rate and it is applicable on a minimal mobile hardware.

AB - Deep learning has revolutionised many fields, but it is still challenging to transfer its success to small mobile robots with minimal hardware. Specifically, some work has been done to this effect in the RoboCup humanoid football domain, but results that are performant and efficient and still generally applicable outside of this domain are lacking. We propose an approach conceptually different from those taken previously. It is based on semantic segmentation and does achieve these desired properties. In detail, it is being able to process full VGA images in real-time on a low-power mobile processor. It can further handle multiple image dimensions without retraining, it does not require specific domain knowledge to achieve a high frame rate and it is applicable on a minimal mobile hardware.

KW - Robotics

KW - Machine Learning

KW - Deep Learning

KW - Computer Vision

KW - Semantic Segmentation

KW - Minimal Hardware

KW - Mobile Robotics

KW - Deep learning

KW - Computer vision

KW - Minimal hardware

KW - Semantic segmentation

KW - Mobile robotics

UR - http://www.scopus.com/inward/record.url?scp=85070709697&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-27544-0_29

DO - 10.1007/978-3-030-27544-0_29

M3 - Conference contribution

SN - 9783030275433

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

SP - 349

EP - 361

BT - RoboCup 2018

A2 - Holz, Dirk

A2 - Genter, Katie

A2 - Saad, Maarouf

A2 - von Stryk, Oskar

PB - Springer Nature

T2 - RoboCup 2018 Symposium

Y2 - 22 June 2018 through 22 June 2018

ER -

Dijk SGV, Scheunemann MM. Deep Learning for Semantic Segmentation on Minimal Hardware. In Holz D, Genter K, Saad M, von Stryk O, editors, RoboCup 2018: Robot World Cup XXII. Springer Nature. 2019. p. 349-361. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-27544-0_29

Deep Learning for Semantic Segmentation on Minimal Hardware

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this