Improving performance on object recognition for real-time on mobile devices

Jin Chun Piao, Hyeon Sub Jung, Chung Pyo Hong, Shin-Dug Kim

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Augmented reality has been on the rise due to the proliferation of mobile devices. At the same time, object recognition has also come to the fore. In particular, many studies have focused on object recognition based on markerless matching. However, most of these studies have focused on desktop systems, which can have high performance in terms of CPU and memory, rather than investigating the use of mobile systems, which have been previously unable to provide high-performance object recognition based on markerless matching. In this paper, we propose a method that uses the OpenCV mobile library to improve real-time object recognition performance on mobile systems. First, we investigate the original object recognition algorithm to identify performance bottlenecks. Second, we optimize the algorithm by analyzing each module and applying appropriate code enhancements. Last, we change the operational structure of the algorithm to improve its performance, changing the execution frequency of the object recognition task from every frame to every four frames for real-time operation. During the three frames in which the original method is not executed, the object is instead recognized using the mobile devices accelerometer. We carry out experiments to reveal how much each aspect of our method improves the overall object recognition performance; overall, experimental performance improves by approximately 800 %, with a corresponding reduction of approximately 1 % in object recognition accuracy. Therefore, the proposed technique can be used to significantly improve the performance of object recognition based on markerless matching on mobile systems for real-time operation.

Original languageEnglish
Pages (from-to)9623-9640
Number of pages18
JournalMultimedia Tools and Applications
Volume75
Issue number16
DOIs
Publication statusPublished - 2016 Aug 1

Fingerprint

Object recognition
Mobile devices
Augmented reality
Accelerometers
Program processors
Data storage equipment

All Science Journal Classification (ASJC) codes

  • Software
  • Media Technology
  • Hardware and Architecture
  • Computer Networks and Communications

Cite this

Piao, Jin Chun ; Jung, Hyeon Sub ; Hong, Chung Pyo ; Kim, Shin-Dug. / Improving performance on object recognition for real-time on mobile devices. In: Multimedia Tools and Applications. 2016 ; Vol. 75, No. 16. pp. 9623-9640.
@article{c7dc5d7e515946b19f657bfae8a3baed,
title = "Improving performance on object recognition for real-time on mobile devices",
abstract = "Augmented reality has been on the rise due to the proliferation of mobile devices. At the same time, object recognition has also come to the fore. In particular, many studies have focused on object recognition based on markerless matching. However, most of these studies have focused on desktop systems, which can have high performance in terms of CPU and memory, rather than investigating the use of mobile systems, which have been previously unable to provide high-performance object recognition based on markerless matching. In this paper, we propose a method that uses the OpenCV mobile library to improve real-time object recognition performance on mobile systems. First, we investigate the original object recognition algorithm to identify performance bottlenecks. Second, we optimize the algorithm by analyzing each module and applying appropriate code enhancements. Last, we change the operational structure of the algorithm to improve its performance, changing the execution frequency of the object recognition task from every frame to every four frames for real-time operation. During the three frames in which the original method is not executed, the object is instead recognized using the mobile devices accelerometer. We carry out experiments to reveal how much each aspect of our method improves the overall object recognition performance; overall, experimental performance improves by approximately 800 {\%}, with a corresponding reduction of approximately 1 {\%} in object recognition accuracy. Therefore, the proposed technique can be used to significantly improve the performance of object recognition based on markerless matching on mobile systems for real-time operation.",
author = "Piao, {Jin Chun} and Jung, {Hyeon Sub} and Hong, {Chung Pyo} and Shin-Dug Kim",
year = "2016",
month = "8",
day = "1",
doi = "10.1007/s11042-015-2999-1",
language = "English",
volume = "75",
pages = "9623--9640",
journal = "Multimedia Tools and Applications",
issn = "1380-7501",
publisher = "Springer Netherlands",
number = "16",

}

Improving performance on object recognition for real-time on mobile devices. / Piao, Jin Chun; Jung, Hyeon Sub; Hong, Chung Pyo; Kim, Shin-Dug.

In: Multimedia Tools and Applications, Vol. 75, No. 16, 01.08.2016, p. 9623-9640.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Improving performance on object recognition for real-time on mobile devices

AU - Piao, Jin Chun

AU - Jung, Hyeon Sub

AU - Hong, Chung Pyo

AU - Kim, Shin-Dug

PY - 2016/8/1

Y1 - 2016/8/1

N2 - Augmented reality has been on the rise due to the proliferation of mobile devices. At the same time, object recognition has also come to the fore. In particular, many studies have focused on object recognition based on markerless matching. However, most of these studies have focused on desktop systems, which can have high performance in terms of CPU and memory, rather than investigating the use of mobile systems, which have been previously unable to provide high-performance object recognition based on markerless matching. In this paper, we propose a method that uses the OpenCV mobile library to improve real-time object recognition performance on mobile systems. First, we investigate the original object recognition algorithm to identify performance bottlenecks. Second, we optimize the algorithm by analyzing each module and applying appropriate code enhancements. Last, we change the operational structure of the algorithm to improve its performance, changing the execution frequency of the object recognition task from every frame to every four frames for real-time operation. During the three frames in which the original method is not executed, the object is instead recognized using the mobile devices accelerometer. We carry out experiments to reveal how much each aspect of our method improves the overall object recognition performance; overall, experimental performance improves by approximately 800 %, with a corresponding reduction of approximately 1 % in object recognition accuracy. Therefore, the proposed technique can be used to significantly improve the performance of object recognition based on markerless matching on mobile systems for real-time operation.

AB - Augmented reality has been on the rise due to the proliferation of mobile devices. At the same time, object recognition has also come to the fore. In particular, many studies have focused on object recognition based on markerless matching. However, most of these studies have focused on desktop systems, which can have high performance in terms of CPU and memory, rather than investigating the use of mobile systems, which have been previously unable to provide high-performance object recognition based on markerless matching. In this paper, we propose a method that uses the OpenCV mobile library to improve real-time object recognition performance on mobile systems. First, we investigate the original object recognition algorithm to identify performance bottlenecks. Second, we optimize the algorithm by analyzing each module and applying appropriate code enhancements. Last, we change the operational structure of the algorithm to improve its performance, changing the execution frequency of the object recognition task from every frame to every four frames for real-time operation. During the three frames in which the original method is not executed, the object is instead recognized using the mobile devices accelerometer. We carry out experiments to reveal how much each aspect of our method improves the overall object recognition performance; overall, experimental performance improves by approximately 800 %, with a corresponding reduction of approximately 1 % in object recognition accuracy. Therefore, the proposed technique can be used to significantly improve the performance of object recognition based on markerless matching on mobile systems for real-time operation.

UR - http://www.scopus.com/inward/record.url?scp=84945566314&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84945566314&partnerID=8YFLogxK

U2 - 10.1007/s11042-015-2999-1

DO - 10.1007/s11042-015-2999-1

M3 - Article

AN - SCOPUS:84945566314

VL - 75

SP - 9623

EP - 9640

JO - Multimedia Tools and Applications

JF - Multimedia Tools and Applications

SN - 1380-7501

IS - 16

ER -