Ocean Eye

https://www.canva.com/design/DAGlwozn7vo/fSDlP3M2W-vYbMh1vFnvlg/edit

Inspiration

According to the CDC, 4500 people each year drowned from 2020 to 2022 in the US, higher than the number in 2019. Drowning is the leading cause of death for children ages 1-4. Less than half of adults have taken swimming lessons, and drowning is more closely associated with minority groups such as African Americans and Hispanics. People, especially young children, are more prone to drowning than ever.

Ocean Eye aims to provide a solution primarily for parents to monitor their children swimming at home. Traditional monitoring tools require the purchase of expensive hardware, such as wearables or underwater sensors. However, Ocean Eye offers a lightweight, AI-powered solution in a common household item, reducing the barrier to entry.

What it does

To begin a swimming session with Ocean Eye, users must confirm their configuration settings. Users are prompted to test their device’s speaker to ensure auditory alerts when a drowning incident is detected. Additionally, users should provide a list of emergency contacts Ocean Eye should send a notification to in the event of an emergency. The user should set up the placement of the camera, ensuring the entire swimming area is in view and is in good quality. Ocean Eye should warn the user if the camera is unable to satisfactorily capture the swimming area due to factors such as lighting or camera angle. The user is able to begin the session. Ocean Eye will now begin to monitor the swimming area. The swimming area is analyzed in real-time for people. Body position and movement is tracked and analyzed to detect potential drowning emergencies. If a drowning incident is detected, Ocean Eye triggers a loud alarm to alert nearby supervisors. Additionally, the provided emergency contacts are sent a notification with a warning of the potential drowning incident. If neither the alarm on the primary device is deactivated, or the emergency contacts react to the alert, emergency services will automatically be contacted to provide assistance.

How we built it

Presentation Layer: Streamlit Powers the delivery of the User Interface components and presentation logic to pass onto the backend for pose detection and video analysis.

Video Analysis: ChatGPT Interprets body position and movement patterns in the water to assess drowning risk. The AI is trained to recognize common drowning behaviors.

Post Detection: Google AI Analyzes body posture to detect common drowning signs such as flailing arms and wrists above shoulders for an extended period of time.

Challenges we ran into

Water reflects light, causing glare and making image detection difficult
Utilizing new technology – Streamlit, AI vision recognition and analysis
API response times – as slow response times can be the difference between life and death
Creating a basic prototype to convey our idea for this project. This is by no means a finished project, and has great potential to continue developing.

What's next for Ocean Eye

Flesh out the configuration settings to ensure optimal environment for alerts and vision analysis
Optimize frame rate to analyze more data for greater accuracy in detecting potential drowning incidents