Research Article
Episodic Reasoning for Vision-Based Human Action Recognition
Maria J. Santofimia,
1
Jesus Martinez-del-Rincon,
2
and Jean-Christophe Nebel
3
1
Computer Architecture and Network Group, School of Computer Science,
University of Castilla-La Mancha, 13072 Ciudad Real, Spain
2
Te Institute of Electronics, Communications and Information Technology (ECIT),
Queens University of Belfast, Belfast BT3 9DT, UK
3
Digital Imaging Research Centre, Kingston University, London KT1 2EE, UK
Correspondence should be addressed to Maria J. Santofmia; mariajose.santofmia@uclm.es
Received 23 August 2013; Accepted 29 October 2013; Published 14 May 2014
Academic Editors: G. Bordogna and I. Garc´ ıa
Copyright © 2014 Maria J. Santofmia et al. Tis is an open access article distributed under the Creative Commons Attribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly
cited.
Smart Spaces, Ambient Intelligence, and Ambient Assisted Living are environmental paradigms that strongly depend on their
capability to recognize human actions. While most solutions rest on sensor value interpretations and video analysis applications,
few have realized the importance of incorporating common-sense capabilities to support the recognition process. Unfortunately,
humanactionrecognitioncannotbesuccessfullyaccomplishedbyonlyanalyzingbodypostures.Onthecontrary,thistaskshould
besupportedbyprofoundknowledgeofhumanagencynatureanditstightconnectiontothereasonsandmotivationsthatexplain
it.Tecombinationofthisknowledgeandtheknowledgeabouthowtheworldworksisessentialforrecognizingandunderstanding
human actions without committing common-senseless mistakes. Tis work demonstrates the impact that episodic reasoning has
in improving the accuracy of a computer vision system for human action recognition. Tis work also presents formalization,
implementation, and evaluation details of the knowledge model that supports the episodic reasoning.
1. Introduction
Recognizing human actions is an essential requirement for
fulflling the vision of Smart Spaces, Ambient Intelligence,
or Ambient Assisted Living. Tese paradigms envision envi-
ronments in which electronic devices, merged with the
background, operate as sensors retrieving environmental
information. Among all diferent types of sensors, video
cameras are extremely powerful devices because of the great
amount of contextual information that they are capable of
capturing. However, despite human’s ability to understand
efortlessly video sequences through observation, computer
vision systems still have work to do in this regard.
Automatic video understanding is a delicate task that yet
remains an unresolved topic [1]. Among all the challenges
involved in video understanding, this paper focuses on
human action recognition since this is an enabling key
for Smart Spaces applications. Applications that depend on
the identifcation of certain behavior require the ability to
recognizeactions.Forexample,kickingandpunchingaretwo
actions that suggest an ongoing fght. In this sense, having
theabilitytorecognizethesequenceofactionsthatdefnean
undesirable behavior can be used to trigger a security alarm.
Obviously, several challenges arise when dealing with
human action recognition. In addition to the inherent
difculty of recognizing diferent people’s body postures
performingthesameaction[2],diferentactionsmayinvolve
similar or identical poses. Moreover, images recorded within
a real environment are not always captured from the best
perspective or angle, which makes it impossible to retrieve
poses consistently [3].
Fortunately, the human ability to recognize actions does
not only rely on visual analysis of human body postures
but also requires additional sources of information such as
context, knowledge about actor intentions, or knowledge
about how the world works normally referred to as common
sense. Tis type of information helps people to recognize,
among several similar actions, the one that is the most
Hindawi Publishing Corporation
e Scientific World Journal
Volume 2014, Article ID 270171, 18 pages
http://dx.doi.org/10.1155/2014/270171