Detection of first and second cardiac sounds based on time frequency analysis

Determining the exact timing o cardiac events, represented by the first (S1) and second (S2) sounds, from the PCG signals (phonocardiogram), represents a great challenge, specially in pathological cases. A system that allows this kind of detection could be used to synchronize several important biomedical devices and diagnosis techniques, such as intraaortic balloons and synchronous images of ultrasound, magnetic resonance and computerized tomography. This work presents a wavelet-based technique for S1 and S2 detection in PCG signals, that is able to perform a good detection in both normal and abnormal cases. It will be shown the criterion of choosing the most suitable wavelet from a set of classical ones, as well as the principles of the final detection method. The results in the analysis of 756 cardiac cycles, from 19 subjects, including normal and abnormal ones present an error ratio of 0.8%, point out to the efficiency of the proposed method.