Towards non-invasive patient monitoring through iris tracking and pain detection

Georges Michael, K. Tsaparellas, G. Panis, C. P. Loizou, A. Lanitis

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Patient monitoring is an important operation taking place in hospitals. It usually involves the use of dedicated invasive equipment that requires the co-operation of patients and also involves remarkable purchase and maintenance costs. In this paper we describe a feasibility study of using image analysis techniques for implementing a low-cost noninvasive patient monitoring system based on iris tracking and pain detection in image sequences captured with ordinary video cameras. Within this context iris tracking can be used for activity monitoring and also as a means for communication in cases where body movement is disabled. Automatic pain detection can be used for detecting increasing pain levels and automatically request help for the patient. As part of our preliminary investigation pain detection is achieved based on a number of texture features extracted from the shape normalized facial regions in image sequences. Iris tracking is carried out by a method based on circular edge detection and isophote curves. The initial results of our study prove the feasibility of the approach as the basis of implementing a complete non-invasive patient monitoring system. Further validation and work in a larger sample of videos is required for further validating the proposed method.

Original languageEnglish
Title of host publicationXIV Mediterranean Conference on Medical and Biological Engineering and Computing, MEDICON 2016
PublisherSpringer Verlag
Pages361-366
Number of pages6
Volume57
ISBN (Print)9783319327013
DOIs
Publication statusPublished - 2016
Event14th Mediterranean Conference on Medical and Biological Engineering and Computing, MEDICON 2016 - Paphos, Cyprus
Duration: 31 Mar 20162 Apr 2016

Other

Other14th Mediterranean Conference on Medical and Biological Engineering and Computing, MEDICON 2016
Country/TerritoryCyprus
CityPaphos
Period31/03/162/04/16

Keywords

  • Iris tracking
  • Pain detection
  • Patient monitoring
  • Texture video analysis

Fingerprint

Dive into the research topics of 'Towards non-invasive patient monitoring through iris tracking and pain detection'. Together they form a unique fingerprint.

Cite this