Ventricular Fibrillation detection

In automatic defibrillation, early detection of the arrhythmia constitutes an essential and extremely sensitive task. Its failure means no shock delivery and, hence, no possible reversal leading to the patient’s death. Besides, as Golden Rule, a shock should not be delivered to a collapsed patient not in cardiac arrest and a successfully defibrillated patient should not be defibrillated again. After defining basic evaluating parameters (sensitivity, specificity, receiver operating curve, positive predictivity and accuracy), several algorithms are reviewed comparing them at the end of the chapter in an attempt to help the designer engineer is his/her selection. Acronyms are used along the text for the sake of space knowing the risk of confusion, although frequently their full identification is repeated and realizing that occasionally the same algorithm shares two abbreviations. To make navigation in this chapter easier, find here listed the seven algorithms treated plus other nine mentioned in the discussion, including also two algorithms for QRS complex detection, and calling attention to some overlapping between the two sets, that is: Probability Density Function (PDF), Threshold Crossing Intervals (TCI), Cardiac Frequency (CF), Signal Morphology (SM) or Correlation Waveform Analysis (CWA), Time-Frequency Analysis (TFA), Wavelet Transform (WT), Phase Space Analysis (PSA), in the first group, followed by Threshold Crossing Intervals (TCI), described in section 4.2.2, AutoCorrelation Fischer (ACF95) algorithm, based on Correlation Waveform Analysis (CWA), explained in section 4.2.4, VF Filter algorithm, after Kuo and Dillman (1978); Spectral (SPEC) algorithm based on Fourier Transform analysis, described in section 4.2.5, Complexity (CPLX) algorithm, the Standard Exponential (STE) algorithm, the Modified Exponential Algorithm (MEA), an STE akin, the Signal Comparison Algorithm (SCA), the Wavelet (WVL) Algorithm, also explained in section 4.2.6. Likewise, two QRS complexes detection algorithms are considered: Tompkins (TOMP, see section 4.2.3), and LI algorithm, (see section 4.2.6).Fil: Laciar Leber, Eric. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - San Juan; ArgentinaFil: Valentinuzzi, Maximo. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina. Universidad de Buenos Aires; Argentina. Universidad Nacional de Tucumán; Argentin