ML Face Alignment Project

This project implements a face alignment system using machine learning techniques in Python. It predicts facial landmarks from images, compares deep learning (using CNNs) and classical machine learning approaches (like Ridge regression), and visualizes results. The implementation is done using PyTorch and scikit-learn in a Jupyter Notebook.

Prerequisites

• Python 3.8 or higher
• Jupyter Notebook

Required Python Packages

Install all dependencies using:

pip install -r requirements.txt

Or, install manually:

pip install numpy torch torchvision scikit-learn matplotlib opencv-python

Data

Place your training and test .npz files in a folder named data:

• data/face_alignment_training_images.npz
• data/face_alignment_test_images.npz

How to Run

1. Activate your virtual environment (if using one):

   python -m venv venv
   

   venv\Scripts\activate  # On Windows
   

   or

   source venv/bin/activate  # On macOS/Linux  
   

2. Install requirements:

   pip install -r requirements.txt
   

3. Start Jupyter Notebook:

   jupyter notebook
   

4. Open the notebook (face_alignment.ipynb) on Jupyter and run all cells.

Output

The notebook will generate a CSV file with predictions (results_task2.csv) in the results folder.

Project Structure

.
├── data/
│   ├── face_alignment_training_images.npz
│   └── face_alignment_test_images.npz
├── results/
│   └── results_task2.csv
├── face_alignment.ipynb
├── requirements.txt
└── README.md

Notes

• Do not upload the venv folder or .ipynb_checkpoints to GitHub.
• For any issues, please open an issue in this repository.