Publications

The ability to detect the sounds S1, S2 and the additional sounds S3 and S4 produced by the heart in its functioning is crucial for the early diagnosis of various cardiac pathologies. The present work proposes a method based on signal power, to localize S1, S2 in the presence of the additional sounds S3 and S4. The proposed method localizes S1 and S2 by highlighting the medium signal energies in which S1 and S2 are contained and considerably reduces the small energies that contain S3 and S4. Next, a comparison of the initial signal with the processed one enables S3 and S4 to be located in the PCG. The proposed method is compared with four methods in the literature. Four PCG recordings containing S3 and S4 respectively, from the Physionet and PASCAL heart sounds challenge databases, gave the proposed method a positive predictive value of 100%, compared with 66.67% for literature methods. This reveals the ability of the new method to localize S1 and S2 in the presence of S3 and S4. Also, the proposed method creates an amplitude difference of 0.171 between S1 and S2, versus 0.071 of the literature methods. This demonstrates the power of the proposed method to differentiate localized S1 from S2. This promising method opens up new prospects for early detection of cardiac abnormalities, improving early intervention and ongoing monitoring of patients suffering from heart disease.

The ability to detect the sounds S1, S2 and the additional sounds S3 and S4 produced by the heart in its functioning is crucial for the early diagnosis of various cardiac pathologies. The present work proposes a method based on signal power, to localize S1, S2 in the presence of the additional sounds S3 and S4. The proposed method localizes S1 and S2 by highlighting the medium signal energies in which S1 and S2 are contained and considerably reduces the small energies that contain S3 and S4. Next, a comparison of the initial signal with the processed one enables S3 and S4 to be located in the PCG. The proposed method is compared with four methods in the literature. Four PCG recordings containing S3 and S4 respectively, from the Physionet and PASCAL heart sounds challenge databases, gave the proposed method a positive predictive value of 100%, compared with 66.67% for literature methods. This reveals the ability of the new method to localize S1 and S2 in the presence of S3 and S4. Also, the proposed method creates an amplitude difference of 0.171 between S1 and S2, versus 0.071 of the literature methods. This demonstrates the power of the proposed method to differentiate localized S1 from S2. This promising method opens up new prospects for early detection of cardiac abnormalities, improving early intervention and ongoing monitoring of patients suffering from heart disease.