# Improving Borderline Adulthood Facial Age Estimation through Ensemble Learning

## TL;DR

This paper introduces an ensemble learning approach combined with a deep learning model to improve age estimation accuracy specifically for borderline adulthood ages, outperforming existing methods and services.

## Contribution

The paper presents a novel ensemble technique that enhances the accuracy of facial age estimation for borderline ages, particularly 16-17 years, using a fine-tuned deep learning model.

## Key findings

- Achieved 68% accuracy for ages 16-17, four times better than DEX.
- Demonstrated improved performance over existing cloud-based and offline services.
- Validated the effectiveness of ensemble learning in borderline age estimation.

## Abstract

Achieving high performance for facial age estimation with subjects in the borderline between adulthood and non-adulthood has always been a challenge. Several studies have used different approaches from the age of a baby to an elder adult and different datasets have been employed to measure the mean absolute error (MAE) ranging between 1.47 to 8 years. The weakness of the algorithms specifically in the borderline has been a motivation for this paper. In our approach, we have developed an ensemble technique that improves the accuracy of underage estimation in conjunction with our deep learning model (DS13K) that has been fine-tuned on the Deep Expectation (DEX) model. We have achieved an accuracy of 68% for the age group 16 to 17 years old, which is 4 times better than the DEX accuracy for such age range. We also present an evaluation of existing cloud-based and offline facial age prediction services, such as Amazon Rekognition, Microsoft Azure Cognitive Services, How-Old.net and DEX.