Title: Movie Genre Classification
Project Description:
I will focus on training neural networks to identify the genre of a movie merely from the elements of it’s poster. The program will be provided a poster for a movie it hasn’t seen before and study it’s various characteristics. The neural network will then choose from a list of various genres and provide a weighted guess.
This neural network will provide an look into the way we as people understand tone given limited information. With only a glance, humans can usually tell what genre of film is being advertised given somewhat minute details. For example, an a famous comedic actor may be a clue that the movie will be a comedy, or dark lighting could signal a drama.
The goal is to train the neural network to act as a human does and quickly decide what the most likely genre of movie is being shown, and gain some insight into the way we percieve ambigous information.
It should be noted that the original idea for this project is to do the same process with the actual films or the scripts, though due to a limited understand of the difficulty of those processes, the idea has been modified to a possibly more managable form.
Project Goals:
- Create a dataset of films and their genres.
- Explore methods of adding noise.
- Train network to identify genres.