XML transcription method for biomedical acoustic signals.

We have proposed applying MIDI technology to coding of biomedical acoustic signals such as heart sounds used in medical diagnosis for constructing medical audio databases. According to our several experiments, we could correspond encoded MIDI event data or musical notes with heart sound components used for clinical diagnosis. However, precise discrimination among abnormal heart sounds and murmurs was difficult owing to insufficiency of analysis precision in our proposed method. Then we have improved its coding precision and supported harmonic or formant analysis in order to extend our application to multimedia low bit-rate audio coding, especially representation of vocal sounds by MIDI codes. In these efforts, we can describe precisely heart-sound components with MIDI codes and segment converted MIDI codes into heart sound components; about 80% automatically in 55 cases. In this paper, we are going to describe our novel heart sound automatic XML transcription method. This XML document includes MIDI data reflecting numerical detailed information for each segmented heart sound component and can reproduce heart sound signals by a browser and MIDI sequencer plug-in.