facial-recognition-system

Table of Contents

• Notebook Recap
• Installation Instructions
  • Dependencies
  • Setup
• Usage
  • Running the Project
  • Example Usage
• Configuration

Notebook Recap

• Posed face recognition as a binary classification problem.
• Implemented one-shot learning for a face recognition problem.
• Applied the triplet loss function to learn a network's parameters in the context of face recognition.
• Mapped face images into 128-dimensional encodings using a pretrained model.
• Performed face verification and face recognition with these encodings.

Installation Instructions

Dependencies

This project requires the following Python libraries and packages:

• tensorflow
• numpy
• pandas
• PIL
• keras

You can install these dependencies using pip:

pip install tensorflow numpy pandas pillow keras

Setup

1. Clone the repository to your local machine:

   git clone https://github.com/justinliu23/facial-recognition-system.git

2. Navigate to the project directory:

   cd face_recognition

3. Ensure all dependencies are installed

Usage

Running the Project

1. Launch Jupyter Notebook:

   jupyter notebook

2. Open the notebook Face_Recognition.ipynb.
3. Execute the cells in the notebook sequentially to run the face recognition pipeline.

Example Usage

Here is a brief guide on how to use the project:

Naive Face Verification

In this section, you will compare two images pixel-by-pixel to determine if they are of the same person.

Encoding Face Images into a 128-Dimensional Vector

Using a Convolutional Neural Network (ConvNet), you will compute encodings for face images.

Using the Triplet Loss

The Triplet Loss function helps learn parameters that improve the accuracy of the model in distinguishing between different faces.

Loading the Pre-trained Model

You will load a pre-trained model to compute the 128-dimensional encodings for the face images.

Applying the Model

• Face Verification: Determine if two images are of the same person (1:1 matching).
• Face Recognition: Identify the person in an image from a set of K persons (1:K matching).

Configuration

Model Configuration

• Input Shape: The input images are expected to have the shape (96, 96, 3).
• ConvNet Architecture: The model architecture includes convolutional layers, batch normalization, pooling layers, and dense layers.
• Loss Function: Triplet loss function is used to train the model.

Data Configuration

• Data Format: The images should be in a format that can be processed by PIL (e.g., JPG, PNG).
• Dataset Structure: Organize your dataset such that each person has their own folder containing images of their face.