Presentation is loading. Please wait.

Presentation is loading. Please wait.

10/23/2015© Mohamed Medhat Gaber1 Adaptive Mobile ECG Analysis Dr Mohamed Medhat Gaber School of Computing University of Portsmouth

Similar presentations


Presentation on theme: "10/23/2015© Mohamed Medhat Gaber1 Adaptive Mobile ECG Analysis Dr Mohamed Medhat Gaber School of Computing University of Portsmouth"— Presentation transcript:

1 10/23/2015© Mohamed Medhat Gaber1 Adaptive Mobile ECG Analysis Dr Mohamed Medhat Gaber School of Computing University of Portsmouth E-mail: mohamed.gaber@port.ac.uk

2 Symbolic Approximation of ECG  Symbolic ApproXimation (SAX) has been invented by Prof. Eamonn Keogh and his colleagues at UCR.  SAX representation of time series is considered the state-of-the-art.  SAX has been successfully used with notable efficiency in: Discord discovery Motif discovery 10/23/2015© Mohamed Medhat Gaber2

3 Real-time Classification of ECG  SAX uses alphabet letters for time series representation.  We have used SAX to represent ECG onboard Google Android for real-time classification using K-NN  We have used adaptation to ensure the continuity of the process 10/23/2015© Mohamed Medhat Gaber3

4 10/23/2015© Mohamed Medhat Gaber4 Current Development  Discord and Motif Discovery Real-time discord and motif discovery of time series Mapping discovered discords and motifs to real ECG problems Identifying other problems that we have termed as regular patters Using SAX and point-based clustering for time series representation

5 10/23/2015© Mohamed Medhat Gaber5 Thanks for listening Dr Mohamed Medhat Gaber School of Computing Faculty of Technology University of Portsmouth E-mail: mohamed.gaber@port.ac.uk, mohamed.m.gaber@gmail.commohamed.gaber@port.ac.ukmohamed.m.gaber@gmail.com Web: http://userweb.port.ac.uk/~gaberm/http://userweb.port.ac.uk/~gaberm/


Download ppt "10/23/2015© Mohamed Medhat Gaber1 Adaptive Mobile ECG Analysis Dr Mohamed Medhat Gaber School of Computing University of Portsmouth"

Similar presentations


Ads by Google