Age Estimator

Estimate the age based on human facial images. To perform the task, we use 3 different machine learning models:

• Multinomial Logistic Regression
• Naïve Bayes classifier
• Multiclass classification with Convolutional Neural Network

Preview

Get Started

1. Ensure all the required dependencies are installed.
2. Initialize the submodule for data scripts. (optional if you've already had it locally)
3. Start the client and server (order doesn't matter).

Install dependencies

• Server

pip install --user --requirement requirements.txt

• Client

cd app

npm i

Initialize data submodule

git submodule update --init

Start the server

In the directory of this repository, do

python -m server.main &

Start the client

cd app

npm run serve &

Data

The datasets is composed of 187154 facial images (> 100,000 samples and > 100 features) for training the model. We combine the following 2 datasets and unify their dimensions into 250x250.

• CACD (160k)
• UTKFace (20k)

Dependencies

• Python 3.7+

• NodeJS 12+

• Python related: Numpy, Pandas, Tensorflow, Pytorch, CV2, PIL

• JS related: Vue, Ramda, Element-UI, Font-Awesome