Face Comparing Project

Introduction

✍ An simple face comparing app using VGG-16 to extract features, compare two photos and see if they are the same person.

Project pipeline:

1. Get input (images) from user.

2. Using MTCNN to detect face.

3. Processing face detected in step 2.

4. Using VGG-16 to extract features as vectors.

5. Evaluating 2 vectors using Cosine Similarity or Euclidean Distance (L2 Norm).

6. Building User Interface (UI) by using Streamlit.

Tutorial

I highly recommend you performing on virtual environment by using Anaconda.

Activating your virtual environment

conda activate [your_venv_name]

Cloning repository

git clone https://github.com/ntvuongg/face-comparing.git

Installing requirements

pip install -r requirements.txt

Running app

streamlit run main.py

Demo