Abstract
This paper presents a new method to represent heart sounds and murmurs in stylized diagrams. Frequency domain analysis, based on an Auto Regressive modelling, shows its superiority against time domain analysis in the identification of the principle sounds (S1 and S2), extra-sounds (S3 and S4) and murmurs. Two separated diagrams give information in the time and frequency domain in order to make a simple representation. The analysis begins by identifying individual cardiac cycles. Then S1, S2, extra-sounds, and murmurs are segmented using spectral estimation. Finally the cycle with the loudest murmur is identified to represent two final diagrams showing its shape in time and frequency content.
| Original language | English |
|---|---|
| Pages (from-to) | 101-104 |
| Number of pages | 4 |
| Journal | Computers in Cardiology |
| State | Published - 1997 |
| Externally published | Yes |
| Event | Proceedings of the 1997 24th Annual Meeting on Computers in Cardiology - Lund, Sweden Duration: 7 Sep 1997 → 10 Sep 1997 |