• 

  overview

• 

  Jad

• 

  James

• 

  Pi

GCI Gauntlet - A gauntlet that uses a Glove - Camera interface with computer vision to change the lives of those with dementia and blindness

💡Inspiration

Millions of people suffer from dementia and blindness. These are horrible conditions that make it difficult to live a normal life, or communicate with loved ones. People with dementia oftentimes don't remember their children's faces, or even their names! Blind people have difficulty communicating and navigating difficult spaces. We wanted to make something that would help those in need and allow them to live a normal life.

🛠️How it works

We used primarily Python and JavaScript. We used a Raspberry Pi 5, Face Recognition Library, OpenCV, and Py3TTSx library to implement vision and TTS. We used HTML CSS JS + A Flask backend to upload videos to the Gauntlet via a separate website. We connected everything together using 3D printed parts, screws, and electrical tape. For hardware, we also used a 3S LiPo battery, a buck converter module, a webcam, and ultrasonic sensor + a Bluetooth connected speaker.

🚧Challenges we faced

We faced quite a few challenges, but in the end we prevailed, and our proud of what we created.

• Difficulty with GPIO Pins + Virtual Environment on the PI
• Integrating the Flask backend
• Server <-> Client Connection

🏆Achievements we made

We successfully made a gauntlet that classifies people's faces based on training data, and a working backend + website that lets you upload videos for training. Blind people and those with dementia can easily use this with wireless earbuds or a speaker to communicate with their loved ones and understand the world around them.

🚀What’s next for the GCI Gauntlet

In the future, we want to improve on this project by improving the client - server connection and allow real time training when you upload videos to the backend. We also want to integrate more sensors, and improve the user experience.

Built With

• css
• face-recognition
• flask
• hardware
• html
• javascript
• opencv
• py3tts
• python
• raspberry-pi