Inspiration
Problem of cyclists who get in accidents without been able to be noticed and saved.
What it does
It takes a POV video of a cyclist ride, analyse it using google cloud videointelligence and transfering it into a text description. Then the video is given to Fetch Ai Text wrapepr Agent, which understand if the cyclist got in an accident. If so we ask a user to identify with their voice if they are fine. If the user says, that they need help - we send a message to an emergency(we emulated it as a TelegramBot)
How we built it
Challenges we ran into
Fetch ai framework. It was really challenging and interesting to get through it. Also it was quite uneasy to deal with computer
Accomplishments that we're proud of
Implimentation of many different frameworks and technology and their communication.
What we learned
How to use OpenAI API, Fetch.ai uagents and how to describer computer vision results with LLM models
What's next for MA3K1
Implement it with better user AI and make it work not just on Windows
Built With
- fetch.ai-uagents
- googlecloudvideointelligence
- openaiapi
- python
- telegrambots
Log in or sign up for Devpost to join the conversation.