College Project Completed – Handwritten Letter OCR

Handwritten Letter OCR

This project is a handwritten letter recognition system using machine learning and OCR techniques.

Dataset

EMNIST (Extended MNIST)

Features

Image preprocessing

Feature extraction using HOG and PCA

Classification using:

Decision Tree

Random Forest

GUI application for drawing or uploading handwritten letters

Confidence score displayed for each model

Technologies Used

Python

OpenCV

scikit-learn

Tkinter

NumPy

How to Run

Install dependencies: pip install -r requirements.txt

Run the application: python main.py

Output

The system predicts the handwritten letter using both models and shows confidence percentages.