Application of a Deep Learning Neural Network for Voiding Dysfunction Diagnosis Using a Vibration Sensor

In a clinical context, there are increasing numbers of people with voiding dysfunction. To date, the methods monitoring status patients have included diary records at home or urodynamic examinations hospitals. The former is less objective and often contains missing data, while latter lacks frequent measurements an invasive procedure. light these shortcomings, this study developed innovative contact-free technique that assists in dysfunction diagnosis. Vibration signals during urination were first detected using accelerometer then converted into mel-frequency cepstrum coefficient (MFCC). Lastly, artificial intelligence model combined uniform manifold approximation projection (UMAP) dimensionality reduction was used to analyze predict six common patterns uroflowmetry assist diagnosing applied database, which data from 76 males aged 30 80 who required for symptoms. resulting system accuracy (precision, recall, f1-score) around 98% both weighted average macro average. This low-cost suitable at-home urinary facilitates long-term uroflow outside hospital checkups. From disease treatment perspective, article also reviews other studies applications intelligence-based monitoring, thus providing helpful diagnostic information physicians.