Final-year-project

Web Application for Automatic Facial Age Estimation of Black Persons(Deep learning approach)

Introduction

Facial age estimation is an important aspect of computer vision with numerous applications, from biometrics to personalized content delivery. This research delves into the domain of Automatic Facial Age Estimation (AFAE) with a dedicated focus on black persons, leveraging a deep learning approach.

Home

Fig1: Web App Home

Tools used

• Programming language and frameworks: Python, Tensorflow, Ktrain, Flask, Git-Bash
• Model training and evaluation: Google colab, Vs code
• WebApp: HTML, CSS, JSON, Python, Flask
• Dataset: BlackFaces
• Offline testing: Vscode terminal and Chrome browser

Flow Chart

Fig2: Flow chart

Project Files

• frontend: contains all Html files associated with the User interface design.
• model: contains the Age prediction model.
• Template: contains the CSS files and bootstrap used to properly render the Html pages for clean and attractive UI.
• images: contains images from project.
• static: contains associated image files used for the AFAE system such as background, favicon.
• test: contains the temporary files used in testing the AFAE system.
• app.py: flask app.
• imageScript.py: Python script used to download the collected images submitted to FAED dataset in batches.
• requirements.txt :To effectively setup and run the web App on a new machine, the user has to install this file.
• storage.py: a python script that captures and stores users’ data in the Storage.json file.
• Storage.json: serves as the Authenticator of user identity as registered or new user before using the AFAE system.
• fyp_age_estimation(ktrain).ipynb: Jupyter notebook used in training the model.
• databank.xlsx: A spreadsheet from which the dataset can be downloaded.
• BlackFaces : contains dataset extracted from databank.xlsx.

How To Use this Web app

1. You can either clone this repo or download the code as zip file.

   • Extract the zip file on your PC.
   • The model size is above Github file size limit. Download the model here and place the model binary files in the model folder
   • Open a Terminal(Cmd/Bash) to run the App

2. It's recommended that a virtual environment is used to install the app packages. Read more on virtual env

   • For Gitbash: Type the code python -m venv afae_venv to setup a virtual environment. And
   • Activate the virtual env. with source venv /afae_venv

3. Install the requirements.txt file to automatically download and setup the app dependencies(packages).

   • Use the command pip install -r requirements.txt to set up the app on your PC

4. Launch the WebApp(Offline Setup)

   • Use python app.py or flask --app run app.py

5. The terminal should display a default IP address to run WebApp in your browser.

   • Ensure the computer is connected to the internet so that the Web UI bootstrap components can be loaded for a beautiful interface.
   • Proceed to test the App

6. Register as new user or loging with initial credentials after your first login.

   • You try the Automatic Facial age estimation as many time as you want.
   • Ensure the system can detect your face(proper room lighting) Fig3: The Age Demo

7. For live testing and Demo anywhere.

   • App not available

   PS: Incase of any bug or misinformation, kindly reach out to the developer of the WebApp via email Taofeek Hammed

Conclusion

Overall, this research contributes valuable insights and a high-performance model and Web Application for Automatic Facial Age Estimation of black persons using a Deep learning approach, with implications for diverse applications in industries such as security, healthcare, and personalized user experiences.

(c) Teekay.ai 2024. All rights reserved.