AIPEDURO

Abbas et al. (2022)

May 15, 2024

Written by:

Penile Curvature, Hypospadias

Article DOI	https://doi.org/10.3389/frai.2022.954497
Objective	To capture for capturing automated measurements of penile curvature based on 2-dimensional images
AI Approach	CNN
Data Source(s)	Nine 3D-printed penile models (900 image dataset)
Model Input	2D image of penis
Model Outcome	Penis curvature
Model Metrics	Mean average Precision = 99.4%, Dice Similarity Coefficient = 98.4% Mean absolute error = 8.5 degrees
Model Usability	NA

AI = Artificial intelligence, CNN = Convolutional neural network