CNN Image Recognition Project: Homer and Bart Simpson

Project Description

This project focuses on developing a Convolutional Neural Network (CNN) to recognize and classify images of Homer and Bart Simpson from the iconic animated television show, "The Simpsons." The objective is to accurately identify characters and potentially extend this to other characters in the franchise.

Objectives

• To build a robust CNN model that effectively identifies Homer and Bart Simpson in various image contexts.
• To explore different neural network architectures and optimization strategies to improve classification accuracy.
• To provide a user-friendly interface for inputting images for character recognition.

Features

• Image upload for character recognition
• Visualization of predictions and probabilities
• Performance metrics reporting (accuracy, precision, recall)
• Support for various image formats (JPEG, PNG)

Technologies Used

• Python
• TensorFlow / Keras
• OpenCV for image processing
• Git for version control
• Jupyter Notebook for experimentation

Setup Instructions

1. Clone the repository:

   git clone https://github.com/DuMilo/cnn-homer-bart.git
   cd cnn-homer-bart

2. Create a virtual environment and activate it:

   python -m venv .venv
   source .venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install the required dependencies:

   pip install -r requirements.txt