Assignment 2 of CS 4501: Vision and Language taught by Prof. Vicente Ordóñez-Román. We will use the MMIMDB dataset (http://lisi1.unal.edu.co/mmimdb/), collected by Arévalo, González, Solorio and Montes-y-Gómez from the Universidad Nacional de Colombia, University of Houston, INAOE-México. This dataset contains data corresponding to ~26,000 movies along with their plots (text) and their movie posters (images), and other information such as ratings and genres (categories). These are more movies than you will watch in a lifetime. We will use this dataset for movie genre classification. Each movie can be labeled with 27 categories of movies.
We will classify the movies into genres
- based on their plots (language)
- based on their posters (vision)
- with a multimodal classifier