Pattern Recognition - Facial Expression Classification

This project explores different machine-learning classifiers to recognize happy and sad faces from grayscale 48×48 facial images. It was developed as part of a Pattern Recognition homework assignment.

The project includes three parts:

Part 1 - Logistic Regression
Part 2 - SVM Classification
Part 3 - CNN Classification

---

Project Structure

Pattern-Recognition-LRnNN/
│
├── .gitignore
├── DeshpandeAryaa_hw2.ipynb        # main notebook 
├── DeshpandeAryaa_hw2.pdf          # final report
│
├── helper.py                       # preprocessing + helper functions
├── Homework2.pdf                   # original assignment sheet
│
├── README.md                       
├── requirements.txt                

---

Setup Instructions

1. Clone the repository

git clone https://github.com/aryaa-deshpande/Pattern-Recognition-LRnNN.git
cd Pattern-Recognition-LRnNN

2. Create a virtual environment

python3 -m venv myenv
source myenv/bin/activate   

3. Install dependencies

pip install -r requirements.txt

---

How to Run

1.	Open the DeshpandeAryaa_hw2.ipynb notebook in Jupyter or VS Code.
2.	Run the cells in order:
    •	Part 1 -> Logistic Regression
    •	Part 2 -> SVM Classification
    •	Part 3 -> CNN Classification
3.	Each section trains a different model and reports:
    •	Validation and test accuracies
    •	Confusion matrix (printed and visualized)

---

Results Summary

Model	Validation Accuracy	Test Accuracy	Key Notes
Logistic Regression	~68%	~71%	Simple linear model; baseline
SVM (Linear Kernel)	~70%	~66%	Similar to logistic regression
SVM (RBF Kernel)	~76%	~74%	Captures nonlinear boundaries
CNN	~89%	~85%	Best performance; learns spatial patterns in faces

---