Inspiration Over the past decade the US has had more than 600 mass shootings every year in schools, malls, and public places. In 2023 there were 656 incidents that killed 754 people and injured 2,443 others, plus 38 school shootings that hurt or killed people. The main problem is that when something scary happens, people just freeze and can't think straight about what to do. That's why I built LifeLine, a safety app that see's what you see and tells you exactly what to do during any emergency like lockdowns, shootings, or medical stuff. When your brain stops working, LifeLine takes over.
What it does LifeLine uses your phone's camera and sensors to scan your environment and gives you step by step instructions during emergencies. It automatically alerts your emergency contacts with your location and what's happening. You can check safety ratings and incident reports for different locations, and you can upload your own incidents so other people know what's happened in different places. There's also breathing exercises to help you stay calm and a smart assistant that can guide you through any emergency situation.
How we built it I built it using SwiftUI for the frontend and Python for the backend. The app uses computer vision to detect objects and assess your surroundings, trained on a huge dataset to recognize as many things as possible. I created decision trees with over 500 different emergency scenarios to give you the right guidance. There's also a conversational system trained on over 1200 different emergency situations that you can talk to for help.
Challenges we ran into The biggest challenge was increasing the backup system's accuracy from 88% to 94.6% which took a lot of fine-tuning. Having Python as the backend work smoothly with the SwiftUI app was really difficult especially for real-time processing. Making sure everything works fast when someone is in an actual emergency was super important and challenging to get right.
Accomplishments I'm really proud that the app can scan a room in just 15 seconds and detect everything important like exits, hazards, and safety equipment. The emergency contact system works perfectly - it immediately sends your location and status to people you trust. The community safety map is really cool too, showing ratings for places like libraries, malls, community centers and every other location with real user reports. The whole thing is simple to use even when you're stressed out.
What I learned I learned a lot about building systems that work under pressure and how important it is to keep things simple when people are scared. I also learned how to combine different technologies like computer vision, location services, and messaging to work together seamlessly. Most importantly I realized how much responsibility comes with making something that could actually save someone's life.
What's next for LifeLine I want to take LifeLine to the next level by using CNN so it guides you through everything, detects how far something is, and makes even more complex decisions like in a thick foggy environment. If you are trying to exit the building, just have LifeLine turned on and listen to what it says. Based on the situation, when LifeLine tells you what to do to stay safe, it will also provide some sort of visual guide so the user has a clear idea of what things should look like.
Log in or sign up for Devpost to join the conversation.