Simpsons Character Recognition
This project will explore multiple CNN models built to identify the Simpons character found in an image file.
The dataset used consists of 20,933 training images belonging to 42 classes. There are an additional 990 images for testing purposes
Models used will range from a no-frills CNN model with no hyperparameter tuning, serving as a baseline model, to a model built using transfer learning with additional tuning of the final layers.
The GitHub repository is available here