Inspiration
We aspired to develop a tool that provides quick, accurate first-aid guidance for injuries. Snapping a picture is much faster than searching for information on Google, by integrating a chatbot the user can get as many details as they need. The AI can even cater information based on their photos to better suit them in their situation.
What it does
Snap AI analyzes photos of wounds or injuries and provides first-aid instructions accordingly based on the image. If the injury is severe, the AI automatically has directions to the nearest hospital and a button for calling emergency services.
How we built it
We developed Snap AI utilizing Flask, HTML, CSS, Python, and JavaScript. For the AI aspect, we utilized Gemini and Roboflow to train and refine our model for accurate classification.
Challenges we ran into
Fine-tuning the AI model's accuracy was arguably one of the major challenges that we faced. Ensuring reliable injury detection required multiple images and time spent in Roboflow. Additionally, styling the CSS design for the Application proved tricky as well.
Accomplishments that we're proud of
We successfully developed a functional AI that can accurately identify wounds and provide first-aid recommendations. There were multiple hurdles and through those we were able to make the AI functional and it has intrinsic real-world value.
What we learned
This project deepened our understanding of AI models and processes to develop them and refine them along with application integration. We also gained valuable experience in problem-solving debugging and AI accuracy.
What's next for Snap AI
We believe that Snap AI has amazing potential. If further refined and perfected it can become a potentially widely-used AI-powered first-aid tool. Future improvement could eventually include expanded injury detection, multilingual support and real-time medical consultations.
Log in or sign up for Devpost to join the conversation.