Biometric Age Estimation Dataset for Minors

10K+ consented face images of minors & young adults aged

10-30 with verified age labels

Check samples on Kaggle

Introduction

The Face Recognition & Age Estimation Dataset is a commercially licensed collection of 10,000+ real phone-captured selfies of individuals aged 10 to 30, with ground-truth age labels verified by birth year. Designed for training accurate age estimation and age-gating models

Built to Solve Real Production Failures

Age estimation models trained on public datasets consistently fail at the most critical boundary — under 18. Three data gaps cause these failures:

Challenge 1: The Under-18 Accuracy Gap

Age estimation errors peak between 13 and 25, exactly where legal compliance demands precision. Public datasets provide little information in the teenage range. This dataset provides dense, per-year coverage from 10 to 30, concentrating data where your model needs it most

Challenge 2: Demographic Bias

MORPH is 80% one ethnicity. IMDB-WIKI skews heavily toward celebrities. Models trained on these datasets produce systematically higher error rates for underrepresented groups. This dataset includes balanced representation across Black, Caucasian, and Hispanic individuals from 5 countries, enabling bias testing and fair model development.

Challenge 3: No Consent, No Commercial Use

UTKFace, FG-NET, and IMDB-WIKI were never consented for commercial deployment. One legal review can block your entire pipeline. Every image in this dataset has full informed consent

Dataset summary

Parameter	Value
Total images	10,000+
Age Range	10-30 years (per-year folders)
Ethnicities	Black, Caucasian, Hispanic
Labels per Image	Age, gender, ethnicity

Production-Ready Quality

10,000+ images organized into per-year age folders (age_10/ through age_30/)
Ground-truth age labels verified by birth year, not estimated or scraped
Multi-ethnic, multi-country participant base for demographic fairness
Clean CSV metadata ready for immediate integration
Ethically collected with full informed consent documentation

Use cases and applications

Age Gating & Compliance. Train models that accurately classify users as under or over 18
Age Estimation & Verification.Build selfie-based age checks that work without ID documents. Verified age labels enable training models while public datasets consistently underperform
Face Recognition.Train age-invariant face recognition models that match individuals across different ages. Multi-year coverage per identity enables robust feature learning despite natural appearance changes

Application Example

Challenge: New legislation requires apps to prevent access for users under 18, but the age estimation model produces unacceptable error rates for teenagers

Solution: Fine-tune the model on dense, per-year labeled data in the 10-30 range with verified ground-truth ages

This Dataset: Real phone selfies of 10-30 year olds with birth-year-verified age labels across multiple ethnicities, targeting exactly where models fail

Download information

A sample version of this dataset is available on Kaggle. Leave a request for additional samples in the form below

Have a question?

Where does the data come from?

We collect data from consented participants recruited across multiple countries. All ages are verified through birth year documentation

What age range does this dataset cover?

The dataset covers ages 10 through 30, organized into per-year folders. This range is specifically designed to cover the legally critical under-18 boundary plus a buffer zone up to 30 for robust model training

Can I use this dataset commercially?

Yes. Unlike public datasets (UTKFace, IMDB-WIKI) which lack commercial consent, this dataset is fully licensed for commercial AI development. Contact us for licensing details

What is the price of the dataset?

The price depends on your specific requirements. Please submit a request to receive a free consultation