An application of wavelet transform for non-invasive evaluation of arteriovenous fistula state

Acoustic signals emitted by the arteriovenous fistula (AVF) can be recorded to assess the AVF condition. This paper concerns the use of wavelet transform for the diagnosis of AVF state. The study were performed using a data set containing samples of the sound signal emitted by the AVF. The aim was to create a solution with multiclass classification. Results show that this approach allows obtaining the high quality information about AVF state. Additionally, the proposed method correlates with the results of the classic diagnostic medical tests.