Classifying ECG signals
A method, including receiving a bipolar signal from a pair of electrodes in proximity to a myocardium of a human subject, and receiving a unipolar signal from a selected one of the pair of electrodes. The method further includes delineating a window of interest (WOI) for the unipolar and bipolar signals, within the WOI computing local unipolar minimum derivatives of the unipolar signal, and times of occurrence of the local unipolar minimum derivatives, and within the WOI computing bipolar derivatives of the bipolar signal at the times of occurrence. The method also includes evaluating ratios of the bipolar derivatives to the local unipolar minimum derivatives, and when the ratios are greater than a preset threshold ratio value, assigning the times of occurrence as times of activation of the myocardium, counting a number of the times of activation; and classifying the unipolar signal according to the number.
1. A method, comprising:
inserting a probe into a heart of a subject, the probe having a pair of electrodes in proximity to a myocardium of a human subject;
providing a processor configured to generate a map of electrical signals in the heart:
receive a bipolar signal with a bipolar feature from the pair of electrodes;
receive a unipolar signal with a unipolar feature from a selected one of the pair of electrodes;
delineate a window of interest (WOI) representative of a time interval for the unipolar and bipolar signals;
within the WOI compute local unipolar minimum derivatives of the unipolar signal and times of occurrence of the local unipolar minimum derivatives;
within the WOI compute bipolar derivatives of the bipolar signal at the times of occurrence;
evaluate ratios of the bipolar derivatives to the local unipolar minimum derivatives;
evaluate the bipolar feature and the unipolar feature using at least one fuzzy function to provide a score representative of a confidence value, wherein the fuzzy function evaluates a ratio of unipolar duration to amplitude of the unipolar signal, and an amplitude of the bipolar signal;
when the ratios are greater than a preset threshold ratio value and the score exceeds a preset threshold score value, assign the times of occurrence as times of activation of the myocardium;
count a number of the times of activation; and
classify the unipolar signal as one selected from the group consisting of a no activation event, a split event, a single activation event and a multiple activation event, according to the number;
providing a filter configured to minimize interferences in the bipolar signal and the unipolar signal prior to computation of the local unipolar minimum derivatives and the bipolar derivatives; and
in a split event, eliminating a weaker feature of a pair of features of the unipolar signal when respective annotations of the pair of features are within a 50 ms time window.
2. The method according to claim 1 , and further comprising, when the bipolar derivatives are less than a preset bipolar derivative threshold, assigning the times of occurrence as the times of activation of the myocardium.
3. The method according to claim 1 , and further comprising, when the local unipolar minimum derivatives are less than a preset unipolar derivative threshold, assigning the times of occurrence as the times of activation of the myocardium.
4. The method according to claim 1 , wherein classifying the unipolar signal comprises defining a plurality of preset classifications for the unipolar signal.
5. The method according to claim 4 , wherein the plurality comprises a first classification wherein the number is zero, a second classification wherein the number is one, a third classification wherein the number is two or three, and a fourth classification wherein the number is greater than three.
6. The method according to claim 1 , and comprising only assigning a given time of occurrence as a given time of activation of the myocardium when a confidence level associated with the given time of occurrence is greater than a preset confidence level.
7. The method according to claim 6 , and comprising only assigning the given time of occurrence as the given time of activation of the myocardium when an amplitude of a corresponding bipolar signal is greater than a preset bipolar signal threshold.
8. The method of claim 1 , wherein the fuzzy function includes a unipolar duration fuzzy function.
9. The method of claim 1 , wherein the fuzzy function includes a unipolar amplitude fuzzy function.
10. The method of claim 1 , wherein the fuzzy function includes a unipolar duration to amplitude ratio fuzzy function.
11. The method of claim 1 , wherein the fuzzy function includes a bipolar amplitude fuzzy function.
12. The method of claim 1 , wherein the fuzzy function is bounded between 0 and 1.
13. The method of claim 1 , wherein the fuzzy function uses:
derivative value of the unipolar signal;
duration of a slope of the unipolar signal;
amplitude of the slope of the unipolar signal at the time window;
a ratio between the duration and the amplitude of the slope of the unipolar signal; and
amplitude of the bipolar signal in the time window.