A System to rate movie based on videos of audience reaction. The goal of this project is to help the movie makers gauge audience reception of the movie, by analysing a small test audience's initial reaction.
We hope that this work will also contribute to creation of smarter audience evaluators, that can gauge the audiences mood and predict which type of content they are likely to enjoy.
- click=7.0
- wxpython=4.0.4
- tensorflow=1.14.0
- matplotlib=3.1.1
- opencv=3.4.2
- tabulate=0.8.3
- tinydb=3.14.1
- python-vlc=3.0.7110 (libVLC required)
keras=2.2.4 scikit-learn=0.21.3 scipy=1.3.1 pandas=0.25.1 pytube=9.5.2 pyserial=3.4 (for gsr sensor)
Given a movie_index.json file present inside "movies" folder, run. python -m scripts.download_movies
Linux/Mac users can simply run run_tests.sh to run all the tests. Otherwise run python -m tests.'test_name' to run a specific test.