Abstract:Background: Age prediction powered by artificial intelligence (AI) can be used as an objective technique to assess the cosmetic effect of rejuvenation surgery. Existing age-estimation models are trained on public datasets with the Caucasian race as the main reference, thus they are impractical for clinical application in Chinese patients. Methods: To develop and select an age-estimation model appropriate for Chinese patients receiving rejuvenation treatment, we obtained a face database of 10 529 images from 1821 patients from the author's hospital and selected two representative age-estimation algorithms for the model training. The prediction accuracies and the interpretability of calculation logic of these two facial age predictors were compared and analyzed. Results: The mean absolute error (MAE) of a traditional support vector machine-learning model was 10.185 years; the proportion of absolute error ≤6 years was 35.90% and 68.50% ≤12 years. The MAE of a deep-learning model based on the VGG-16 framework was 3.011 years; the proportion of absolute error ≤6 years was 90.20% and 100% ≤12 years. Compared with deep learning, traditional machine-learning models have clearer computational logic, which allows them to give clinicians more specific treatment recommendations. Conclusion: Experimental results show that deep-learning exceeds traditional machine learning in the prediction of Chinese cosmetic patients' age. Although traditional machine learning model has better interpretability than deep-learning model, deep-learning is more accurate for clinical quantitative evaluation. Knowing the decision-making logic behind the accurate prediction of deep-learning is crucial for deeper clinical application, and requires further exploration.

Face Age Prediction Based on Machine Learning

Face Prediction Model for an Automatic Age-invariant Face Recognition System

Automatic Age Estimation Based on Deep Learning Algorithm.

An Intelligent Human Age Prediction from Face Image Framework Based on Deep Learning Algorithms

Research on the Modeling of Fast Face Recognition Against Age Disturbance under Deep Learning

Age Classification with Deep Learning Face Representation

Human Face Recognition and Age Estimation with Machine Learning

Aging Face Recognition: A Hierarchical Learning Model Based on Local Patterns Selection

Exploring artificial intelligence from a clinical perspective: A comparison and application analysis of two facial age predictors trained on a large-scale Chinese cosmetic patient database

Facial Age Estimation with Age Difference

Age Prediction From Face Images Via Contrastive Learning

Age transformation based on deep learning: a survey

Apparent age prediction from faces: A survey of modern approaches

Learning from facial aging patterns for automatic age estimation

Image-Based Human Age Estimation by Manifold Learning and Locally Adjusted Robust Regression

Deep-learning based descriptors in application to aging problem in face recognition

Deep Probabilities for Age Estimation

Age Estimation Guided Convolutional Neural Network for Age-Invariant Face Recognition

Facial age estimation by multilinear subspace analysis

Can estimate age range using'a face a person'?

Facial Age Estimation Using Bio-Inspired Features and Cost-Sensitive Ordinal Hyperplane Rank