Handwritten Digit Recognition Project

This project focuses on recognizing handwritten digits (0–9) using a K-Nearest Neighbors (KNN) classifier trained on a custom dataset.
It allows users to draw a digit on a canvas and predicts the drawn digit using the trained model.

📂 Project Structure

• Dataset:

  • 28x28 pixel grayscale images.
  • Organized in folders named 0, 1, ..., 9.
  • Approximately 999 images per digit.

• Model:

  • K-Nearest Neighbors (KNN) classifier.
  • Trained on the flattened pixel data.
  • Model is saved locally for future predictions.

• User Interface:

  • A simple canvas in Google Colab for users to draw digits.
  • Captures and preprocesses the drawing to match model input format.

🚀 How to Run

1. Prepare the Dataset:

   • (Provided in the repo)
   • Ensure images are in the correct folder structure (0–9).
   • Each image should be 28x28 pixels.

2. Train the Model:

   • Run the training code to fit the KNN model.
   • The model will be saved after training.

3. Test the Model:

   • Draw a digit on the canvas provided in the notebook.
   • The model will predict and display the drawn digit.

🛠 Technologies Used

• Python 3
• OpenCV
• NumPy
• Scikit-learn (KNeighborsClassifier)
• Google Colab

📋 Requirements

Install the following packages if running locally:

pip install opencv-python scikit-learn numpy matplotlib

For Quick testing

Load the collab file

Load the models

Run the collab cells for loading and testing | exclude the dataset