Robust and Accurate Anomaly Detection in ECG Artifacts Using Time Series Motif Discovery

Algorithm 2

Robust anomaly detection.

Input: ECG in lead ,

     : The starting position of P-wave of th beat,

     : The end position of P-wave of th beat,

      : The position of Q point of th beat,

     : The end position of T-wave of th beat,

numberofbeat: The number of R

cananomaly: Anomaly candidates from the cleanest lead.

Output: All anomaly beats

(1) For to numberofBeat  //Begins beat extraction

(2)   = Extract()  //P-wave of th beat

(3)   = Extract()  //PR segment of th beat

(4)   = Extract()  //QT interval of th beat

(5) End

(6) For to numberofBeat

(7)  nearestneighbordis = infinity

(8)  For to numberofBeat

(9)   If !=

(10)    distance = DTW(, ) +

(11)         DTW(, ) +

(12)         DTW(, )

(13)    If distance < nearestneighbordis[]

(14)     nearestneighbordis[] = distance

(15)    End

(16)   End

(17)  End

(18) End

(19) newanomaly = Find (nearestneighbordis > thres)

(20) anomalybeats = Merge_answer(Shift(cananomaly), newanomaly)

(21) Return anomalybeats