Elektrokardiyografik işaretlerde QRS deteksiyon algoritmaları
Başlık çevirisi mevcut değil.
- Tez No: 75441
- Danışmanlar: DOÇ. DR. MEHMET KORÜREK
- Tez Türü: Yüksek Lisans
- Konular: Elektrik ve Elektronik Mühendisliği, Electrical and Electronics Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 1998
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: Elektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
- Bilim Dalı: Biyomedikal Mühendisliği Bilim Dalı
- Sayfa Sayısı: 78
Özet
Bu tezde, EKG işaretinin oluşumu, normal EKG işareti, EKG işaretindeki bozukluklar, EKG işaretinin algılanması konularının ardından bu işaretin bilgisayar yardımı ile incelenmesini sağlayacak olan algoritmalar incelenecektir. EKG işaretindeki ritm ve şekil bozuklukları incelenebilecek ve bu bozuklukları bilgisayar yardımı ile bulabilecek programların yazılmasına çalışılacaktır. EKG işaretindeki ritm ve şekil bozukluklarına aritmi adı verilir. Aritmi hastalan sürekli gözetim altında tutulan ve yoğun hemşire bakımı isteyen hastalardır. Bu yüzden hastalara bağlanan aritmi dedektörleri sayesinde yoğun bakım ünitelerinde bir rahatlama görülmüştür. Hastanın sürekli olarak incelenmesi sırasında EKG kullanılmaktadır. Böylece hastanın günlük kayıtlan tutulmakta ve ani ölüm riski azaltılabilmektedir. Günümüzde tam anlamıyla mükemmel olmayan bilgisayar teşhisleri uzman bir doktor tarafından gözden geçirilerek teşhise varılmaktadır. Gene de bilgisayar tarafından yapılan EKG incelemeleri fiyatlan düşürmekte ve daha geniş kullanıma yol açmaktadır.
Özet (Çeviri)
The anxions data related heart from ECG records is obtained by two types studies; ECG shape corruptions and rhytim corruptions. But, to obtain these data, We need long time (more than one day) ECG records. To seek shape corruptoins lack of a program control in ECG records take long time. To detect QRS complex from ECG signal is diffticult, not only this signals physiologic variations, at the same time, different noise types. The muscle movement as a corruption affect, electrode shifting, mixing of network frequency and noise delivered from base line variation are examples of this. Digital filters used in this study removed this aftects. The first of QRS complex detection algorithms in this study contains the main operations, digital filters, non-linear transformations and decision making algorithm. In the other four algoritms, 'AF'letters represented amplitude and the first derivation algorithms,“FD”letters represent only the first derivation algorithms,“FS”letters represent the first and the second derivation algorithms, and the final category“DF”letters represent algorithms explained each of relative noise sensitivity response of informed four types of noise by the use of MIT-BIH data base. ECG signals that used in diagnonis have required to clear from noise. The amplitude of ECG signal varies between 10 mv - 5 mv and have a frequency range of about 0,02 - 150 Hz. The signals in these intervals have been easily affected from biological and external signals.Figure (1) shows ECG records that have been affected various noises First method,Detection of QRS complex 1st methode, Dedection of QRS complex Figure (2) shows the block diagram which used to detect QRS complex. The base- line alternating affect having about 0.3 Hz frequency have an important rule in detecting QRS complex. Using a high - pass filter which have 2Hz cut- off frequency, base- line alteration is removed on ECG signal. There are two sub parts of this digital filter output.Because of the frequency spectrum of QRS complex is like 17 Hz a band pass filter. İt passed from a BPF whose central frequency is 17 Hz. Though the energy of signal is increased band pass filter output is passed through square wave block and was obvious to the other signals. The differatiation equation of this operation. y(k) = x2(k) (l)(a) (b) (c) Figure:l Noisy electrocardiograms a)muscle contraction b)baseline drift and respration c)power line and motion artifact BPF SQR INT HPF LPF DLY 12 T MULT Figure :2. Block diagram of QRS detection vuThe next operation is moving window integration.İntegration block and two peaks that wan presented are combiued and to be a positive signal. The differentiation equation for this integration is shown below, y(k)=(l/N).[x (k-(N-l))+x (k-(N-2)) +.... + x (k)] (2) In this equation, N is the sampling number which surraunds the integration width.The other output of HPF is passed 36 Hz cut-off frequency could be removed on signal. A delay of 12 T was put on the output of the block.The arm of this operation (delay operation) is to compansate the delays coming from filter and required for the program is, y(n)=x(n-12) (3) The output of delay block and the output of block which have been increased the energy of QRS signal are multiplied with mutiplier circuit. So the noise on ECG signal can be eliminated by the above operation and detect the place of QRS. Digital Filters High Pass Filter We based on analog (S domain Butterwort 2n4 derivention)low pass filter for this H = 2 *, (4) s2 + 1.41s + l To filter baseline alteration, we select 2Hz cut-off frequency. This transfer functon explained to obtain the correlations between digital and analog filters. W s->^ (5) Wa= Analog cut-off frequency of Low Pass Filter In this sub episode explained to design firstly high pass filter and secondly BPF Then passing z domain from s domain by using bilineer transformation After We obtained the transfer function that used in digital filter and from this the differen tiation equations. Low-Pass Filters In design based on Z domain to prevent a missing of maximum frequency on f spectrum we sample with a signal of F = -£- If we think the maximum frequency 2 of signal is 150 Hz we sample with frequency of signal to be 300 Hz.If we select the vmfrequency of LPF 36 Hz. Because of sample frequency 360 Hz, k is equal to 10.(k=10) So we obtained transfer function and from this the derivation equations. y (k) = 2y (k-1) - y(k-2) + u(k) - 2 u (k-20) + u (k-20) (6) Second method,Algorithms Based on Amplitude and First Derivative (AF) The concept for this algorithm was taken from Gustafson [7]. The first derivative is calculated at each point of the ECG: Y(n) = X(n+1) - X(n-l) 1< n < 8190 (7) The first derivate array is then searched for points which exceed a constant threshold: Y(i) > 0.15. (8) Then the next three derivative values Y(i+1), Y(i+2), and Y(i+3) must also exceed 0.15. If the above conditions are met, point i can be classified as a QRS candidate if the next two sample points have positive slope amplitude products: Y(i+l)X(i+l) and Y(i+2) X(i+2) > 0. (9) Third methodjAIgorithms Based on First Derivative Only (FD) This algorithm was adapted from one developed by Menard [7]. The first derivative is calculated for each point of the ECG, using the formula specified by Menard: Y (n) = - 2 X (n-2) - X (n-1) + X (n+1) - X (n+1) + 2X(n+2) 2 Slope threshold (12) IXFourth methodcAIgorithms Based on First and Second Derivative (FS) This algorihm is a simplification of the QRS detection scheme presented by Balda [7]. The absolute values of the first and second derivative are calculated from the ECG: YO (n) = ABS [X(n+1)- X (n-1)] 21.0 (16) Once this occurs, the next eight points are compared the threshold. If six or more of these eight points meet exceed the threshold, the criteria for identification of a QRS candidate is met. Fifth methode, Algorhitms Based on Digital Filters (DF) This algorithm is an adaptation of Okada s QRS detection algorithm [7]. The first stage smooths the ECG using a three-point moving average filter: YO (n) = [ X ( n-1) + 2 X (n) + X(n+1)] / 4 1< n< 8190 ( 1 7) The output of the moving point averaging filter is passed through a low-pass filter. n+m Yl (n) = [1 / (2m +1)] ^YO(k) m 0 (21) Y4 (n) = 0 otherwise (22) The maximum value of this array is determined and scaled to form the threshold : Threshold =0.125 max [Y4 (n)] mthreshold then,QRS candidate.
Benzer Tezler
- Sözdizimi metodu ile elektrokardiyografi işaretlerinin analizi
Başlık çevirisi yok
SERGİN ÖZENÇ
Yüksek Lisans
Türkçe
1996
Elektrik ve Elektronik Mühendisliğiİstanbul Teknik ÜniversitesiDOÇ.DR. MEHMET KÜREK
- Elektrokardiyografi işaretlerini gerçek zamanda bilgisayara aktaran ve işleyen arayüz programı tasarımı
Developing an user interface program to transfer and process ECG data
ALİ NİZAM
Yüksek Lisans
Türkçe
2000
Elektrik ve Elektronik Mühendisliğiİstanbul Teknik ÜniversitesiDOÇ.DR. MEHMET KORUEK
- Dalgacık ağlarıyla elektrokardiyografik aritmilerin sınıflandırılması
Electrocardiogram arrhytmias classification using wavelet networks
ABDULLAH KARADAĞ
Yüksek Lisans
Türkçe
2009
Biyomühendislikİstanbul Teknik ÜniversitesiElektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
DOÇ. DR. MEHMET KORÜREK
- Makine öğrenmesi yaklaşımları kullanılarak EKG işaretlerinden aritmi teşhisi
Diagnosis of arrhythmia from ECG signs using machine learning approaches
ERSİN ERSOY
Doktora
Türkçe
2022
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve KontrolAnkara ÜniversitesiBilgisayar Mühendisliği Ana Bilim Dalı
DOÇ. DR. GAZİ ERKAN BOSTANCI
- Hilbert transformundan yararlanarak EKG işaretlerinin analizi ve sınıflanması
Hilbert transform to ECG signal analysis and classification
HAKKI ANIL ŞEKER
Yüksek Lisans
Türkçe
2002
Elektrik ve Elektronik Mühendisliğiİstanbul Teknik ÜniversitesiMimarlık Ana Bilim Dalı
DOÇ. DR. MEHMET KORÜREK