Real-Time Handwritten Digit Recognition:

1. Introduction:

This project uses a Logistic Regression model trained on the MNIST dataset to recognize handwritten digits (0–9) in real-time using a webcam feed. It combines machine learning and computer vision techniques to make live predictions from camera input.

2. Tech Stack:

• Python
• OpenCV
• NumPy
• Pandas
• scikit-learn
• Pillow (PIL)
• Matplotlib

3. Model Details:

• Dataset: MNIST (70,000 grayscale images of handwritten digits).
• Model Used: Logistic Regression (multinomial).
• Accuracy: ~90–92% on test data.

4. Instructions:

1. A camera window will open.
2. Write a digit (0–9) on paper and hold it inside the green box.
3. The predicted digit will appear in the console.
4. Press ‘q’ to quit.