Circapproved: Digital Pattern Recognition via Artificial Neural Network for the Identification of Normal Penis...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Urology

Background:

• Circumcision requires anatomical assessment for eligibility, necessitating identification of contraindications like congenital anomalies.
• Current assessment methods can be subjective, highlighting the need for objective, reliable tools.
• Penile anatomy evaluation is crucial for determining suitability for circumcision.

Purpose of the Study:

• To develop and validate an AI-based digital pattern recognition system for assessing penile anatomy.
• To create a mobile application for accurate identification of normal penile parameters for circumcision eligibility.
• To enhance the precision and efficiency of the circumcision approval process.

Main Methods:

• An Artificial Neural Network (ANN) was trained using digital images of penile anatomy from dorsal, lateral, and ventral views.
• Image augmentation and transfer learning techniques were employed to address dataset limitations.
• Ethical guidelines, including informed consent, were followed throughout the study.

Main Results:

• The AI model, integrated into the Circapproved app, achieved high accuracy: 88.89% (dorsal), 90.91% (lateral), and 92.5% (ventral).
• The system demonstrated potential for large-scale screening and streamlining circumcision approvals.
• A cloud-based deployment ensured scalability and accessibility of the AI model.

Conclusions:

• AI integration can significantly improve the accuracy of the circumcision approval process.
• The developed AI-driven mobile application shows promise for widespread use, particularly in resource-limited settings.
• Further dataset expansion is recommended to enhance AI model accuracy.