Getting started

Installation

The Video Opinion Mining(VEM) module can be downloaded either from pip or github, with the respective commands.

  • pip install VideoOpinionMining

  • git+https://github.com/Labic-ICMC-USP/VEM.git

To use, it is also needed to install some other modules. Most of them are listed in the file requirements.txt, so running pip install -r requirements.txt and later using the following commands should be enough to get everything you need to run.

git clone https://github.com/m3hrdadfi/soxan
mv soxan/*

In case you are using colab, there is an example with all the modules being installed in the link https://colab.research.google.com/drive/1UqzA6bDgtZWGji652a0UjT7ZtDh3Xyi5?authuser=1#scrollTo=d3QJ6iSd-o3D

First steps

After installing the module with all of it’s dependencies, import the vem module.

from VideoOpinionMining import vem

Then create an instance of the class VEMProcessor passing the path to the video you want to be analyzed as a parameter.

video_file = "emotions.mp4"
example = vem.VEMProcessor(video_file)

Now, with a single command you can run all the analysis

The results can then be seen either in the dataframe returned, or in the heatmaps generated in a folder called heatmaps.

Instead of a single line command, you can also use other classes to make your analysis your own way, more details and extra uses can be found at the documentations session.