A Branch-and-Bound Framework for Unsupervised Common Event Discovery

Wen Sheng Chu, Fernando de la Torre, Jeffrey F. Cohn, Daniel S Messinger

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Event discovery aims to discover a temporal segment of interest, such as human behavior, actions or activities. Most approaches to event discovery within or between time series use supervised learning. This becomes problematic when relevant event labels are unknown, are difficult to detect, or not all possible combinations of events have been anticipated. To overcome these problems, this paper explores Common Event Discovery (CED), a new problem that aims to discover common events of variable-length segments in an unsupervised manner. A potential solution to CED is searching over all possible pairs of segments, which would incur a prohibitive quartic cost. In this paper, we propose an efficient branch-and-bound (B&B) framework that avoids exhaustive search while guaranteeing a globally optimal solution. To this end, we derive novel bounding functions for various commonality measures and provide extensions to multiple commonality discovery and accelerated search. The B&B framework takes as input any multidimensional signal that can be quantified into histograms. A generalization of the framework can be readily applied to discover events at the same or different times (synchrony and event commonality, respectively). We consider extensions to video search and supervised event detection. The effectiveness of the B&B framework is evaluated in motion capture of deliberate behavior and in video of spontaneous facial behavior in diverse interpersonal contexts: interviews, small groups of young adults, and parent-infant face-to-face interaction.

Original languageEnglish (US)
Pages (from-to)1-20
Number of pages20
JournalInternational Journal of Computer Vision
DOIs
StateAccepted/In press - Feb 9 2017

Fingerprint

Supervised learning
Labels
Time series
Costs

Keywords

  • Bag-of-words
  • Branch and bound
  • Common event discovery
  • Event detection
  • Global optimization
  • Human interaction
  • Synchrony discovery
  • Unsupervised learning
  • Video indexing

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this

A Branch-and-Bound Framework for Unsupervised Common Event Discovery. / Chu, Wen Sheng; de la Torre, Fernando; Cohn, Jeffrey F.; Messinger, Daniel S.

In: International Journal of Computer Vision, 09.02.2017, p. 1-20.

Research output: Contribution to journalArticle

@article{669851c0d1cc48beae06aa82de4a66fd,
title = "A Branch-and-Bound Framework for Unsupervised Common Event Discovery",
abstract = "Event discovery aims to discover a temporal segment of interest, such as human behavior, actions or activities. Most approaches to event discovery within or between time series use supervised learning. This becomes problematic when relevant event labels are unknown, are difficult to detect, or not all possible combinations of events have been anticipated. To overcome these problems, this paper explores Common Event Discovery (CED), a new problem that aims to discover common events of variable-length segments in an unsupervised manner. A potential solution to CED is searching over all possible pairs of segments, which would incur a prohibitive quartic cost. In this paper, we propose an efficient branch-and-bound (B&B) framework that avoids exhaustive search while guaranteeing a globally optimal solution. To this end, we derive novel bounding functions for various commonality measures and provide extensions to multiple commonality discovery and accelerated search. The B&B framework takes as input any multidimensional signal that can be quantified into histograms. A generalization of the framework can be readily applied to discover events at the same or different times (synchrony and event commonality, respectively). We consider extensions to video search and supervised event detection. The effectiveness of the B&B framework is evaluated in motion capture of deliberate behavior and in video of spontaneous facial behavior in diverse interpersonal contexts: interviews, small groups of young adults, and parent-infant face-to-face interaction.",
keywords = "Bag-of-words, Branch and bound, Common event discovery, Event detection, Global optimization, Human interaction, Synchrony discovery, Unsupervised learning, Video indexing",
author = "Chu, {Wen Sheng} and {de la Torre}, Fernando and Cohn, {Jeffrey F.} and Messinger, {Daniel S}",
year = "2017",
month = "2",
day = "9",
doi = "10.1007/s11263-017-0989-7",
language = "English (US)",
pages = "1--20",
journal = "International Journal of Computer Vision",
issn = "0920-5691",
publisher = "Springer Netherlands",

}

TY - JOUR

T1 - A Branch-and-Bound Framework for Unsupervised Common Event Discovery

AU - Chu, Wen Sheng

AU - de la Torre, Fernando

AU - Cohn, Jeffrey F.

AU - Messinger, Daniel S

PY - 2017/2/9

Y1 - 2017/2/9

N2 - Event discovery aims to discover a temporal segment of interest, such as human behavior, actions or activities. Most approaches to event discovery within or between time series use supervised learning. This becomes problematic when relevant event labels are unknown, are difficult to detect, or not all possible combinations of events have been anticipated. To overcome these problems, this paper explores Common Event Discovery (CED), a new problem that aims to discover common events of variable-length segments in an unsupervised manner. A potential solution to CED is searching over all possible pairs of segments, which would incur a prohibitive quartic cost. In this paper, we propose an efficient branch-and-bound (B&B) framework that avoids exhaustive search while guaranteeing a globally optimal solution. To this end, we derive novel bounding functions for various commonality measures and provide extensions to multiple commonality discovery and accelerated search. The B&B framework takes as input any multidimensional signal that can be quantified into histograms. A generalization of the framework can be readily applied to discover events at the same or different times (synchrony and event commonality, respectively). We consider extensions to video search and supervised event detection. The effectiveness of the B&B framework is evaluated in motion capture of deliberate behavior and in video of spontaneous facial behavior in diverse interpersonal contexts: interviews, small groups of young adults, and parent-infant face-to-face interaction.

AB - Event discovery aims to discover a temporal segment of interest, such as human behavior, actions or activities. Most approaches to event discovery within or between time series use supervised learning. This becomes problematic when relevant event labels are unknown, are difficult to detect, or not all possible combinations of events have been anticipated. To overcome these problems, this paper explores Common Event Discovery (CED), a new problem that aims to discover common events of variable-length segments in an unsupervised manner. A potential solution to CED is searching over all possible pairs of segments, which would incur a prohibitive quartic cost. In this paper, we propose an efficient branch-and-bound (B&B) framework that avoids exhaustive search while guaranteeing a globally optimal solution. To this end, we derive novel bounding functions for various commonality measures and provide extensions to multiple commonality discovery and accelerated search. The B&B framework takes as input any multidimensional signal that can be quantified into histograms. A generalization of the framework can be readily applied to discover events at the same or different times (synchrony and event commonality, respectively). We consider extensions to video search and supervised event detection. The effectiveness of the B&B framework is evaluated in motion capture of deliberate behavior and in video of spontaneous facial behavior in diverse interpersonal contexts: interviews, small groups of young adults, and parent-infant face-to-face interaction.

KW - Bag-of-words

KW - Branch and bound

KW - Common event discovery

KW - Event detection

KW - Global optimization

KW - Human interaction

KW - Synchrony discovery

KW - Unsupervised learning

KW - Video indexing

UR - http://www.scopus.com/inward/record.url?scp=85011835144&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85011835144&partnerID=8YFLogxK

U2 - 10.1007/s11263-017-0989-7

DO - 10.1007/s11263-017-0989-7

M3 - Article

AN - SCOPUS:85011835144

SP - 1

EP - 20

JO - International Journal of Computer Vision

JF - International Journal of Computer Vision

SN - 0920-5691

ER -