Uroflowmetry measures the flow of urine and tracks its flow rate, the voiding volume and total voiding time. Therefore, uroflowmetry remains an important tool for the assessment of patients with lower urinary tract symptoms (LUTS). Due to the time-consuming nature of urologist appointments and the resulting sparse acquisition of flow data, its long-term diagnostic potential is very limited. Acoustic voiding testing (Sonouroflowmetry) appears to correlate well with conventional uroflowmetry and is a promising alternative for home urinary flow monitoring.

In order to exploit the potential of home monitoring by sonouroflowmetry, training data was acquired under artificial and real voiding conditions. Within the scope of the thesis, features should be extracted with the help of an optimized Mel filter bank, thus allowing to classify the different flow rates. In addition, a data augmentation should be implemented that allows characteristics of real flow rates to be transferred to artificial flow rates.