Handwritten Digit Recognizer (LeNet-5 CNN)

Interactive LeNet-5: A Classic CNN in Action

Step into the world of deep learning with this unique, interactive project that showcases the LeNet-5 convolutional neural network (CNN) in real-time. Created entirely in the LÖVE 2D framework, this demo lets you draw a number with your mouse and watch a fully-fledged LeNet-5 model attempt to recognize your handwriting.

LeNet-5 is a landmark in AI history. Unlike a simple feedforward network, it uses convolutional layers to automatically learn spatial hierarchies of features, making it incredibly robust at recognizing visual patterns. This demo provides a rare glimpse into its inner workings, allowing you to see this classic architecture process your input from raw pixels to a final prediction.

This is a complete implementation of the LeNet-5 architecture, optimized to run efficiently on the client side. The model was trained on the MNIST dataset with data augmentation techniques to ensure it adapts well to different handwriting styles.

How to Use:

• Left Mouse Button: Draw on the canvas.
• Right Mouse Button: Erase what you've drawn.
• Hold Middle Mouse Button: Drag to move the canvas.
• Press 'C': Clear the entire screen.
• Press 'Esc': Quit the application.

Experience the power of a foundational deep learning model firsthand!