Skip to content Skip to navigation

Complex Event Recognition for Activity Recognition

The goal in activity recognition is to recognise activities taking place between entities, e.g., people meeting or moving together, by exploiting information about observed activities of individuals. We have been developing a complex event recognition system for activity recognition in the Event Calculus that uses video content as its main source of information.

Recognition and learning methods

We perform complex event recognition using the ‘Event Calculus for Run-Time reasoning’ (RTEC), an open-source Prolog implementation of the Event Calculus, designed to compute continuous temporal projection queries for pattern matching on data streams. Moreover, we have developed methods for learning these Event Calculus patterns from data. Our methods include fast online learning algorithms such as OLED, as well as learning algorithms that can learn patterns for probabilistic event recognition, such as OSLα and wOLED. We are also investigating semi-supervised learning methods, such as SPLICE, for completing the missing supervision, thus enabling our existing learners to operate as if the training sequence was fully supervised.

Detected activities

We detect:

  • meeting (when two people are meeting),
  • moving (when two people are moving together),
  • fighting (when two or more people are fighting),
  • leaving object (when someone leaves an object unattended).

Publications

Michelioudakis E., Artikis A. and Paliouras G., Semi-Supervised Online Structure Learning for Composite Event Recognition, Machine Learning, pp. 1-22, 2019. (pdf)

Katzouris N, Michelioudakis E., Artikis A. and Paliouras G. Online Learning of Weighted Relational Rules for Complex Event Recognition, European Conference on Machine Learning (ECML-PKDD) 2018. (pdf)

Katzouris N., Artikis A., Paliouras G.: Online Learning of Event Definitions, Theory and Practice of Logic Programming (TPLP), special issue for the 32nd International Conference of Logic Programming (ICLP 2016). (pdf)

Micheloudakis V., Skarlatidis A., Paliouras G and Artikis A: OSLa: Online Structure Learning using Background Knowledge Axiomatization, European Conference on Machine Learning (ECML-PKDD) 2016 (pdf)

Artikis A., Sergot M. and Paliouras G. An Event Calculus for Event Recognition. IEEE Transactions on Knowledge and Data Engineering (TKDE), 27(4): 895-908.(pdf)

Artikis A., Sergot M. and Paliouras G. A Logic Programming Approach to Activity Recognition, ACM International Workshop on Events in Multimedia, 2010. (pdf)