Fig. 1
Workflow of a convolutional neural network (CNN) applied to panoramic radiographs for the estimation of sex and age. A total of 9349 radiographs from individuals aged from 6 to 22.99 years old (y. o.) were used. Images, sex, and chronological age were provided as input features. Data were split into five folds for cross-validation, with training (blue) and validation (yellow) sets. Image preprocessing was applied prior to CNN feature extraction through convolution and pooling layers. Fully connected layers integrated extracted features to predict two outcomes: sex (male [♂] vs. female [♀]) and age (> 15 years old vs. ≤ 15 years old).
