In this paper, we proposed comprehensive voiding analysis and management system, initially focused on identifying dysuria and subsequent strictures. However, we intend to expand it sufficiently to become an open system in the future to implement a comprehensive, personalized, artificial intelligence-based support system that can identify, correlate, and manage a variety of voiding disorders based on biometric data, including pathological voiding disorders that can be identified and treated with endoscopic surgery, as well as pathological voiding disorders that can be identified and treated with endoscopic surgery. The goal is to enable full-cycle management of a single patient from prediction, diagnosis, treatment, and management.

The current proposed system consists of a monitoring system and an intraoperative identification system. The first part identifies the patient's posture and determines the time of voiding based on the gyro-accelerometer data provided by the smart wristband. A short- and long-term memory-based network is used to segment characteristic accelerometer sequences and identify the time of voiding, which is similar to a 24-hour Holter monitor for urination. The second part of the system involves an endoscopic imaging system that can identify urethral strictures in the field based on the ResNet 50 architecture. Further enhancements to this system could include an augmented reality (AR) overlay. By combining these two systems, a diagnostic application pairing can be created that can match voiding patterns with pathophysiologic visual images of the urethra.

Accuracy was based on established clinical guidelines: monitoring accuracy correctly classified urination times 95.8% of the time based on the movement of the smart wristband alone. Meanwhile, it had an average sensitivity of 0.96 when identifying strictures for surgery.

Accuracy was based on established clinical guidelines: monitoring accuracy correctly classified urination times 95.8% of the time based on the movement of the smart wristband alone. Meanwhile, it had an average sensitivity of 0.96 when identifying strictures for surgery.

The potential of the proposed system demonstrates the potential for point-of-care surgical applications that can provide highly accurate and easy-to-use diagnostic tools and AR-based surgical guidance. Together, these systems can provide a truly comprehensive, lifestyle-based, integrated diagnostic surgical overlay system that can identify and treat pathologic urethral conditions. In the future, we look forward to being able to manage the entire lifecycle of a patient's urologic condition.

Greff K, Srivastava RK, Koutnik J, Steunebrink BR, Schmidhuber J. LSTM: a search space odyssey. IEEE Trans Neural Netw Learn Syst 2017;28:2222-32.
Wen L, Li X, Gao L. A transfer convolutional neural network for fault diagnosis based on ResNet-50. Neural Comput Appl 2020;32: 6111-24.