Transcript Classification of Electrocardiogram (ECG) Waveforms for the

Classification of Electrocardiogram
(ECG) Waveforms for the Detection of
Cardiac Problems
By Enda Moloney
Contents









Project aims
The Heart & ECG
ECG Signals
MIT-BIH arrhythmia Database
QRS Detection
Pan-Tomkins Algorithm
Artificial Neural Network
Timeline
Question
Project Aims







Analyse ECG waveform to detect abnormalities
Using sample waveform from MIT-BIH database
Process waveforms to make it easier to classify them
Extract information from ECG waves e.g. QRS
complex
Use the artificial Neural Network to classify the ECG
waves into different classes
Translate the ECG classification system from Matlab
to C
Possible development of a suitable of
hardware/software system and Database
Heart & ECG




Determining if the heart is
performing normally or
suffering from abnormalities
e.g. skipped heartbeats.
Indicating previous damage
to the heart muscle.
Providing information on the
physical condition of the
heart.
Been used to detect noncardiac diseases
ECG Signal

An ECG is measuring the
electrical potential between
various points of the body
using leads. The normal
ECG wave is composed of




The P wave
QRS complex
The T wave
The relationship between P
waves and QRS complexes
helps distinguish various
cardiac irregularities.
MIT-BIH Arrhythmia Database

This is a waveform of
the data 101 of the MITBIH Database of ECG
waveforms
Using matlab
programme is used to
extract information for
the MIT-BIH arrhythmia
database.
ECG signal 101.dat
2
1.5
1
Voltage / mV

0.5
0 28
1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111
-0.5
-1
0
10
20
30
40
Time / s
50
60
70
80
ECG signal 101.dat
ECG signal 101.dat
2
1.5
Voltage / mV
1
0.5
0 28
1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111
-0.5
-1
0
10
20
30
40
Time / s
50
60
70
80
QRS Detection


The QRS complex is the
most important complex in
the ECG. The duration and
amplitude sure be measure
as accurate as possible.
There are two methods:

the Pan-Tompkins
algorithm

the derivation-based
method.
Pan-Tompkins algorithm

Pan-Tompkins algorithm proposes a real-time QRS detection
algorithm based on slope, amplitude and width of the QRS
complexes
After Squaring

0.18
0.16
0.14
0.12
0.1

0.08
0.06
0.04
0.02
0
0
0.5
1
1.5
2
2.5
3
4
x 10
After the implementing
the Bandpass filter &
differentiation this
suppresses P and T
waves.
Squaring makes all the
results positive and
emphasising from large
differences arising for
the QRS complexes
Artificial Neural Network


ANN is an adaptive system
that changes its structure
based on external or internal
information that flows through
the network during the
learning phase
When the ECG waves have
been processed, they must
be classified into Two classes


Normal
Abnormal
Timeline

30-Jan-09


15-Feb-09


Transfer the ECG system from Matlab to C, as a
real-time Implementation. The neural network
needs to be in C.
Develop hardware circuit to interact with the
software, thus a circuit that has a ECG sensor
9-Mar-09

Investigate possible extensions of the system. Eg.
Web-based database system that could be used
story cardiology records and analysis
Question