Joint segmentation and temporal structure inference for partially-observed event sequences

Harvey Thornburg, Dilip Swaminathan, Todd Ingalls, Randal Leistikow

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

3 Citations (Scopus)

Abstract

Many events of interest in human activity-based multimedia applications exhibit a high degree of temporal structure. This structure generates expectancies regarding the occurrence and location of subsequent events. In the context of switching state-space models, we develop a general Bayesian framework for representing temporal expectancies and fusing them with raw sense-data to improve both event segmentation and temporal structure identification. Furthermore, we develop a new cognitive model for event anticipation which adapts to incoming sense-data in real time. Comparative advantages of the proposed framework are realized in controlled experiments involving partially-observed, quasi-periodic event streams.

Original languageEnglish (US)
Title of host publication2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006
PublisherIEEE Computer Society
Pages41-44
Number of pages4
ISBN (Print)0780397517, 9780780397514
DOIs
StatePublished - Jan 1 2006
Event2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006 - Victoria, BC, Canada
Duration: Oct 3 2006Oct 6 2006

Publication series

Name2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006

Other

Other2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006
CountryCanada
CityVictoria, BC
Period10/3/0610/6/06

Fingerprint

Identification (control systems)
Experiments

ASJC Scopus subject areas

  • Signal Processing

Cite this

Thornburg, H., Swaminathan, D., Ingalls, T., & Leistikow, R. (2006). Joint segmentation and temporal structure inference for partially-observed event sequences. In 2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006 (pp. 41-44). [4064515] (2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006). IEEE Computer Society. https://doi.org/10.1109/MMSP.2006.285265

Joint segmentation and temporal structure inference for partially-observed event sequences. / Thornburg, Harvey; Swaminathan, Dilip; Ingalls, Todd; Leistikow, Randal.

2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006. IEEE Computer Society, 2006. p. 41-44 4064515 (2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006).

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

Thornburg, H, Swaminathan, D, Ingalls, T & Leistikow, R 2006, Joint segmentation and temporal structure inference for partially-observed event sequences. in 2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006., 4064515, 2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006, IEEE Computer Society, pp. 41-44, 2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006, Victoria, BC, Canada, 10/3/06. https://doi.org/10.1109/MMSP.2006.285265
Thornburg H, Swaminathan D, Ingalls T, Leistikow R. Joint segmentation and temporal structure inference for partially-observed event sequences. In 2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006. IEEE Computer Society. 2006. p. 41-44. 4064515. (2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006). https://doi.org/10.1109/MMSP.2006.285265
Thornburg, Harvey ; Swaminathan, Dilip ; Ingalls, Todd ; Leistikow, Randal. / Joint segmentation and temporal structure inference for partially-observed event sequences. 2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006. IEEE Computer Society, 2006. pp. 41-44 (2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006).
@inproceedings{a59a2c82ed4f467994a950bfc22b182b,
title = "Joint segmentation and temporal structure inference for partially-observed event sequences",
abstract = "Many events of interest in human activity-based multimedia applications exhibit a high degree of temporal structure. This structure generates expectancies regarding the occurrence and location of subsequent events. In the context of switching state-space models, we develop a general Bayesian framework for representing temporal expectancies and fusing them with raw sense-data to improve both event segmentation and temporal structure identification. Furthermore, we develop a new cognitive model for event anticipation which adapts to incoming sense-data in real time. Comparative advantages of the proposed framework are realized in controlled experiments involving partially-observed, quasi-periodic event streams.",
author = "Harvey Thornburg and Dilip Swaminathan and Todd Ingalls and Randal Leistikow",
year = "2006",
month = "1",
day = "1",
doi = "10.1109/MMSP.2006.285265",
language = "English (US)",
isbn = "0780397517",
series = "2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006",
publisher = "IEEE Computer Society",
pages = "41--44",
booktitle = "2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006",

}

TY - GEN

T1 - Joint segmentation and temporal structure inference for partially-observed event sequences

AU - Thornburg, Harvey

AU - Swaminathan, Dilip

AU - Ingalls, Todd

AU - Leistikow, Randal

PY - 2006/1/1

Y1 - 2006/1/1

N2 - Many events of interest in human activity-based multimedia applications exhibit a high degree of temporal structure. This structure generates expectancies regarding the occurrence and location of subsequent events. In the context of switching state-space models, we develop a general Bayesian framework for representing temporal expectancies and fusing them with raw sense-data to improve both event segmentation and temporal structure identification. Furthermore, we develop a new cognitive model for event anticipation which adapts to incoming sense-data in real time. Comparative advantages of the proposed framework are realized in controlled experiments involving partially-observed, quasi-periodic event streams.

AB - Many events of interest in human activity-based multimedia applications exhibit a high degree of temporal structure. This structure generates expectancies regarding the occurrence and location of subsequent events. In the context of switching state-space models, we develop a general Bayesian framework for representing temporal expectancies and fusing them with raw sense-data to improve both event segmentation and temporal structure identification. Furthermore, we develop a new cognitive model for event anticipation which adapts to incoming sense-data in real time. Comparative advantages of the proposed framework are realized in controlled experiments involving partially-observed, quasi-periodic event streams.

UR - http://www.scopus.com/inward/record.url?scp=34250739338&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34250739338&partnerID=8YFLogxK

U2 - 10.1109/MMSP.2006.285265

DO - 10.1109/MMSP.2006.285265

M3 - Conference contribution

AN - SCOPUS:34250739338

SN - 0780397517

SN - 9780780397514

T3 - 2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006

SP - 41

EP - 44

BT - 2006 IEEE 8th Workshop on Multimedia Signal Processing, MMSP 2006

PB - IEEE Computer Society

ER -