nathandpenha/European-Space-Agency-Video-Sentiment-Analysis

Deep space missions will pose new challenges to astronauts’ health and under such extreme conditions, preserving a healthy emotional state will be one of the main hurdles. Thus being able to assess the emotional state of astronauts in these situations will be key in guaranteeing the completion of the mission. The past decade has also witnessed the rise of Artificial Intelligence, with the improvements on sentiment analysis of particular relevance to this activity. While sentiment analysis has been traditionally applied only to text data due to recent advancements in machine learning (ML), deep learning (DL) and computer vision (CV) are also starting to be applied to audio and images/video data. These new approaches open the possibility of developing a facial and voice sentiment analysis system that could be deployed on-board of the spacecraft to assess in real time the emotional states of the astronauts

Video Sentiment Analysis

Repository for the Video sentiment analysis module of the ESA project.

Setup the project

1. Clone/pull project

• If it is from scratch, clone this project:

git clone https://ooti-projects.win.tue.nl/gitlab/st-c2019/esa/video-sentimental-analysis.git

• if it is from existing project, pull the project

git pull

2. Follow the instructions to install DVC (on Windows):

• Linux: https://dvc.org/doc/install/linux
• MacOS: https://dvc.org/doc/install/macos
• Windows: https://dvc.org/doc/install/windows

3. Follow the instructions to install aws cli version 2.

• Linux: https://docs.aws.amazon.com/cli/latest/userguide/install-cliv2-linux.html
• MacOS: https://docs.aws.amazon.com/cli/latest/userguide/install-cliv2-mac.html
• Windows: https://docs.aws.amazon.com/cli/latest/userguide/install-cliv2-windows.html

4. Configure aws cli:

> aws configure
AWS Access Key ID [None]: AKIA55GEKVQR3FCPU7PV
AWS Secret Access Key [None]: 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
Default region name [None]: eu-central-1
Default output format [None]: text

5. Download production data and models:

dvc pull

Use cases

Modify production data

1. Add or remove audios from the prod_data directory.
2. Execute dvc status to see that the contents of prod_data were modified.
3. Execute dvc add prod_data to update the contents of prod_data.dvc.
4. Track changes with git: git add prod_data.dvc.
5. Git commit: git commit -m "commit message" .
6. Update production: dvc push.
7. Push changes to GitLab: git push origin [branch-name].

Add new production model

1. Add new model in the prod_models directory.
2. Execute dvc status to see that the contents of prod_models were modified.
3. Execute dvc add prod_models to update the contents of prod_models.dvc.
4. Track changes with git: git add prod_models.dvc.
5. Git commit: git commit -m "commit message" .
6. Update production: dvc push.
7. Push changes to GitLab: git push origin [branch-name].

Deploy new model in the raspberry pi

1. Log into the raspberry pi.
2. Go to /home/pi/esaProject.
3. Activate virtual environment: source videoTeam/bin/activate.
4. Go to the repository directorycd video-sentiment-analysis.
5. Pull changes from master git pull origin master.
6. Pull latest data and models using dvc dvc pull.
7. Go to the repository directory containing the source code.
8. Run your script.