Inspiration
As new parents, the anxiety of ensuring the safety and well-being of our infants can be overwhelming. Traditional baby monitors, which offer basic sound and video, often fall short of providing actionable insights that can prevent tragedies like Sudden Infant Death Syndrome (SIDS) and accidents. Inspired by the need for a more intelligent and proactive solution, we developed Safenest.AI – a state-of-the-art baby monitor that leverages advanced AI technologies to safeguard our precious little ones.
What it does
SafeNest AI is an innovative baby monitoring system designed to detect potential hazards and monitor the baby’s well-being in real-time. Here’s how it works:
• Hazard Detection: Using AI-powered surveillance cameras and sensors, Safenest.AI continuously monitors the baby’s environment for potential hazards. The system can identify dangerous objects, unsafe sleeping positions, and other risks.
• Emotion and Health Monitoring: Integrating HumeAI, Safenest.AI reads the baby’s emotions and stress levels. It tracks vital signs such as breathing, heart rate, and body temperature to provide a comprehensive overview of the baby’s health.
• Data Analysis and Alerts: The data collected is processed using AWS Bedrock and the Claude 3.5 API to analyze the baby’s behavior and environment. When a high danger level is detected, Safenest.AI sends instant alerts and notifications to parents via SMS and a mobile app.
• Data Storage and Insights: All data is securely stored in SQL storage and used to generate detailed dashboards and summaries, helping parents gain valuable insights into their baby’s habits and well-being.
How we built it
Frontend
We developed an intuitive user interface where parents can view real-time video feeds, health metrics, and receive alerts with react user interface. The interface allows easy access to historical data and insights, providing a user-friendly experience.
Backend
We built a distributed system connecting to a hypothetical vitals server with the following implementations:
• Data Collection: Surveillance cameras and wearable sensors on the baby collect continuous data, including video footage and biometric readings.
• Data Processing: The collected data is preprocessed using OpenCV for efficiency, then processed in real-time using AWS Bedrock and Claude 3.5 API. HumeAI is employed to assess the baby’s emotional state and stress levels.
• Risk Assessment: The processed data is analyzed to determine the danger level of the baby and the environment. If a high danger level is detected, the system triggers notifications and SMS alerts to parents.
• Storage and Dashboard: All data is stored securely in SQL storage, ensuring easy retrieval and compliance with data privacy regulations. The stored data is used to generate comprehensive dashboards that provide parents with actionable insights.
Technology Stack
• Surveillance Cameras and Sensors: For continuous data collection.
• AWS Bedrock, OpenCV and Claude 3.5 API: For data processing and risk assessment.
• HumeAI: For emotion and stress level detection.
• SQL Storage: For secure data storage and retrieval.
• Twilio: For sending SMS notifications.
• Mobile App and Dashboard: For real-time monitoring and insights.
• React: For building user interfaces in the mobile app and dashboard.
Challenges we ran into
• Data Privacy and Security: Ensuring compliance with HIPAA and GDPR regulations and implementing strong security protocols, including end-to-end encryption, to protect user data.
• High API Costs: Managing the high costs associated with live data streaming and API usage by optimizing and conditioning API triggers.
• Integration Difficulties: Integrating HumeAI and AWS Bedrock was particularly challenging since this was our first time working with these technologies. We had to learn their APIs and find effective ways to combine their outputs for a seamless user experience.
Accomplishments that we're proud of
• Successfully developing a fully functional prototype of Safenest.AI with advanced hazard detection and health monitoring capabilities.
• Ensuring data privacy and security while providing parents with peace of mind through real-time alerts and insights.
• Creating a user-friendly interface that empowers parents with actionable data and comprehensive monitoring of their baby’s well-being.
• Leveraging multiple different LLMs and servers for a thorough and accurate analysis of real-time data.
What we learned
Through the development of Safenest.AI, we gained valuable insights into:
• Advanced AI technologies and their application in real-time hazard detection and health monitoring.
• The importance of data privacy and security in developing health-related technologies.
• Effective integration of various technologies to create a seamless and reliable product.
What's next for Safenest.AI
We plan to explore expanding into nursery and childcare centers and create rules to limit the API pull to reduce costs. Additionally, we aim to build an actual model that can process video footage, examine the qualification of the lowest video resolution for faster generation, and set rules for the number of frames per time to lower the data usage while ensuring the model's performance.
Log in or sign up for Devpost to join the conversation.