AIPEDURO

Abbas et al. (2023)

May 15, 2024

Written by:

Penile curvature, Hypospadias

Article DOI	https://doi.org/10.3389/fped.2023.1149318
Objective	To determine accurate penile curvature without requiring arc length
AI Approach	CNN
Data Source(s)	Nine 3D-printed penile models (913 image dataset); 4 intraoperative penile images
Model Input	Image of hypospadias
Model Outcome	Penile Curvature
Model Metrics	Abs. mean average error < 5 degrees AI variance (for cases of 30 deg curv) = 1.7 deg AI variance (for cases of 70 deg curv) = 6 deg
Model Usability	NA

AI = Artificial Intelligence, CNN = Convolutional neural network