Physical environment virtualization for human activities recognition

Azin Poshtkar, Vinayak Elangovan, Amir Shirkhodaie, Alex Chan, Shuowen Hu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

Human activity recognition research relies heavily on extensive datasets to verify and validate performance of activity recognition algorithms. However, obtaining real datasets are expensive and highly time consuming. A physics-based virtual simulation can accelerate the development of context based human activity recognition algorithms and techniques by generating relevant training and testing videos simulating diverse operational scenarios. In this paper, we discuss in detail the requisite capabilities of a virtual environment to aid as a test bed for evaluating and enhancing activity recognition algorithms. To demonstrate the numerous advantages of virtual environment development, a newly developed virtual environment simulation modeling (VESM) environment is presented here to generate calibrated multisource imagery datasets suitable for development and testing of recognition algorithms for context-based human activities. The VESM environment serves as a versatile test bed to generate a vast amount of realistic data for training and testing of sensor processing algorithms. To demonstrate the effectiveness of VESM environment, we present various simulated scenarios and processed results to infer proper semantic annotations from the high fidelity imagery data for human-vehicle activity recognition under different operational contexts.

Original languageEnglish (US)
Title of host publicationModeling and Simulation for Defense Systems and Applications X
EditorsEric J. Kelmelis
PublisherSPIE
ISBN (Electronic)9781628415940
DOIs
StatePublished - Jan 1 2015
EventModeling and Simulation for Defense Systems and Applications X - Baltimore, United States
Duration: Apr 21 2015 → …

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Volume9478
ISSN (Print)0277-786X
ISSN (Electronic)1996-756X

Other

OtherModeling and Simulation for Defense Systems and Applications X
CountryUnited States
CityBaltimore
Period4/21/15 → …

Fingerprint

Activity Recognition
Virtualization
Virtual Environments
Virtual reality
Recognition Algorithm
Simulation Modeling
environment simulation
Testbed
Testing
test stands
Computer simulation
imagery
education
Scenarios
Semantic Annotation
annotations
semantics
Fidelity
Demonstrate
Accelerate

All Science Journal Classification (ASJC) codes

  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics
  • Computer Science Applications
  • Applied Mathematics
  • Electrical and Electronic Engineering

Cite this

Poshtkar, A., Elangovan, V., Shirkhodaie, A., Chan, A., & Hu, S. (2015). Physical environment virtualization for human activities recognition. In E. J. Kelmelis (Ed.), Modeling and Simulation for Defense Systems and Applications X [94780I] (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 9478). SPIE. https://doi.org/10.1117/12.2178547
Poshtkar, Azin ; Elangovan, Vinayak ; Shirkhodaie, Amir ; Chan, Alex ; Hu, Shuowen. / Physical environment virtualization for human activities recognition. Modeling and Simulation for Defense Systems and Applications X. editor / Eric J. Kelmelis. SPIE, 2015. (Proceedings of SPIE - The International Society for Optical Engineering).
@inproceedings{3dd5a10bc0984c639dc477adb8b7ddb2,
title = "Physical environment virtualization for human activities recognition",
abstract = "Human activity recognition research relies heavily on extensive datasets to verify and validate performance of activity recognition algorithms. However, obtaining real datasets are expensive and highly time consuming. A physics-based virtual simulation can accelerate the development of context based human activity recognition algorithms and techniques by generating relevant training and testing videos simulating diverse operational scenarios. In this paper, we discuss in detail the requisite capabilities of a virtual environment to aid as a test bed for evaluating and enhancing activity recognition algorithms. To demonstrate the numerous advantages of virtual environment development, a newly developed virtual environment simulation modeling (VESM) environment is presented here to generate calibrated multisource imagery datasets suitable for development and testing of recognition algorithms for context-based human activities. The VESM environment serves as a versatile test bed to generate a vast amount of realistic data for training and testing of sensor processing algorithms. To demonstrate the effectiveness of VESM environment, we present various simulated scenarios and processed results to infer proper semantic annotations from the high fidelity imagery data for human-vehicle activity recognition under different operational contexts.",
author = "Azin Poshtkar and Vinayak Elangovan and Amir Shirkhodaie and Alex Chan and Shuowen Hu",
year = "2015",
month = "1",
day = "1",
doi = "10.1117/12.2178547",
language = "English (US)",
series = "Proceedings of SPIE - The International Society for Optical Engineering",
publisher = "SPIE",
editor = "Kelmelis, {Eric J.}",
booktitle = "Modeling and Simulation for Defense Systems and Applications X",
address = "United States",

}

Poshtkar, A, Elangovan, V, Shirkhodaie, A, Chan, A & Hu, S 2015, Physical environment virtualization for human activities recognition. in EJ Kelmelis (ed.), Modeling and Simulation for Defense Systems and Applications X., 94780I, Proceedings of SPIE - The International Society for Optical Engineering, vol. 9478, SPIE, Modeling and Simulation for Defense Systems and Applications X, Baltimore, United States, 4/21/15. https://doi.org/10.1117/12.2178547

Physical environment virtualization for human activities recognition. / Poshtkar, Azin; Elangovan, Vinayak; Shirkhodaie, Amir; Chan, Alex; Hu, Shuowen.

Modeling and Simulation for Defense Systems and Applications X. ed. / Eric J. Kelmelis. SPIE, 2015. 94780I (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 9478).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Physical environment virtualization for human activities recognition

AU - Poshtkar, Azin

AU - Elangovan, Vinayak

AU - Shirkhodaie, Amir

AU - Chan, Alex

AU - Hu, Shuowen

PY - 2015/1/1

Y1 - 2015/1/1

N2 - Human activity recognition research relies heavily on extensive datasets to verify and validate performance of activity recognition algorithms. However, obtaining real datasets are expensive and highly time consuming. A physics-based virtual simulation can accelerate the development of context based human activity recognition algorithms and techniques by generating relevant training and testing videos simulating diverse operational scenarios. In this paper, we discuss in detail the requisite capabilities of a virtual environment to aid as a test bed for evaluating and enhancing activity recognition algorithms. To demonstrate the numerous advantages of virtual environment development, a newly developed virtual environment simulation modeling (VESM) environment is presented here to generate calibrated multisource imagery datasets suitable for development and testing of recognition algorithms for context-based human activities. The VESM environment serves as a versatile test bed to generate a vast amount of realistic data for training and testing of sensor processing algorithms. To demonstrate the effectiveness of VESM environment, we present various simulated scenarios and processed results to infer proper semantic annotations from the high fidelity imagery data for human-vehicle activity recognition under different operational contexts.

AB - Human activity recognition research relies heavily on extensive datasets to verify and validate performance of activity recognition algorithms. However, obtaining real datasets are expensive and highly time consuming. A physics-based virtual simulation can accelerate the development of context based human activity recognition algorithms and techniques by generating relevant training and testing videos simulating diverse operational scenarios. In this paper, we discuss in detail the requisite capabilities of a virtual environment to aid as a test bed for evaluating and enhancing activity recognition algorithms. To demonstrate the numerous advantages of virtual environment development, a newly developed virtual environment simulation modeling (VESM) environment is presented here to generate calibrated multisource imagery datasets suitable for development and testing of recognition algorithms for context-based human activities. The VESM environment serves as a versatile test bed to generate a vast amount of realistic data for training and testing of sensor processing algorithms. To demonstrate the effectiveness of VESM environment, we present various simulated scenarios and processed results to infer proper semantic annotations from the high fidelity imagery data for human-vehicle activity recognition under different operational contexts.

UR - http://www.scopus.com/inward/record.url?scp=84943578448&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84943578448&partnerID=8YFLogxK

U2 - 10.1117/12.2178547

DO - 10.1117/12.2178547

M3 - Conference contribution

AN - SCOPUS:84943578448

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Modeling and Simulation for Defense Systems and Applications X

A2 - Kelmelis, Eric J.

PB - SPIE

ER -

Poshtkar A, Elangovan V, Shirkhodaie A, Chan A, Hu S. Physical environment virtualization for human activities recognition. In Kelmelis EJ, editor, Modeling and Simulation for Defense Systems and Applications X. SPIE. 2015. 94780I. (Proceedings of SPIE - The International Society for Optical Engineering). https://doi.org/10.1117/12.2178547