Military and Refugee Sentiment Classifier

Inspiration

Determine emotional and mental health of military personnel during and after service by identifying personnel in need of support. It helps in identification of PTSD.

It is also used to recognize refugees in need of special treatment, support, and supplies.

What it does

Input: A set of images of refugees or military forces/veterans Output: The most prevalent emotion among the set, Frequency of each emotion (displayed in data table), Each image’s sentiment/emotion

How we built it

Python - Used to develop the model Tensorflow Keras Sequential Model - Convolutional Neural Network Flask - Creates connection between HTML & Python HTML/CSS/Javascript - Used to create the website

Challenges we ran into

Finding an adequate dataset to train on Connecting frontend (HTML) to backend (Python) - solved using Flask

Accomplishments that we're proud of

Created a complete product that can have a large impact on the military personnel and refugees throughout the world, thereby improving their lives.

What we learned

How to use the Flask package How to create a robust website How to create an emotion detector using AI/ML

What's next for Military and Refugee Sentiment Classifier

Image Recognition - Identify people who may need support Database - store abnormalities in people's patterns Improve accuracy to 95%+

Built With

• ai
• css
• flask
• html
• javascript
• keras
• ml
• neural-networks
• python
• tensorflow