SleepSync

About

Our project aims to transform sleep health management by leveraging advanced technology. Combining ReactJS for frontend and Flask, NodeJS, and MongoDB for backend, we've crafted a versatile platform facilitating virtual consultations with healthcare professionals, with a specialized emphasis on sleep health. Through machine learning algorithms, we provide tailored doctor recommendations based on user symptoms, ensuring individualized care across a spectrum of health concerns.

A pivotal aspect of our platform is integrating EEG data analysis from wearable devices, enabling precise sleep stage classification via convolutional neural networks. This innovative approach identifies sleep irregularities and extends to broader health issues, enhancing the platform's utility and impact.

Navigating challenges such as interpreting intricate research and integrating real-time consultation features, our project embodies the convergence of healthcare and technology. Our comprehensive solution stands at the forefront of healthcare innovation, offering a transformative approach to managing diverse health issues with virtual consultations and data-driven insights.

Link

https://youtu.be/QXVtFZ0E9u4

Inspiration

The inspiration behind our project stemmed from a desire to tackle the pervasive issue of sleep disorders and the growing need for accessible healthcare solutions. Recognizing the potential of technology to bridge gaps in healthcare access, we were motivated to develop a platform that could empower individuals to understand better and manage their sleep health.

How we built it

The project was developed iteratively, commencing with the design of the user-friendly front end using ReactJS. A combination of Flask, NodeJS, and ExpressJS was employed for the backend to establish resilient APIs. MongoDB was selected as the database solution, ensuring efficient data storage and retrieval. The integration of Flask enriched the backend infrastructure, enhancing its flexibility and scalability.

Incorporating GraphQL into the system facilitated optimized data communication, resulting in improved bandwidth utilization and enabling seamless interactions between components. Additionally, the project featured EEG data classification using Convolutional Neural Networks (CNNs). This advanced technique allowed for precise EEG data analysis obtained from wearable devices, enabling accurate sleep stage classification.

Challenges we ran into

One of the primary challenges was deciphering complex research papers on sleep stage classification and EEG data analysis. Understanding and implementing convolutional neural networks for segmenting EEG data into sleep stages required meticulous experimentation and fine-tuning. Additionally, integrating real-time virtual consultations using sockets and WebRTC posed technical hurdles that demanded innovative solutions.

Accomplishments that we're proud of

We're proud to achieve an 85% accuracy and 77% F1 score after training our Convolutional Neural Network (CNN) for EEG data analysis. This accomplishment underscores the effectiveness of our software, which integrates EEG data to classify sleep stages accurately. Our platform, built with ReactJS for frontend and Flask, NodeJS, ExpressJS, and MongoDB for backend, offers intuitive virtual consultations and personalized doctor recommendations based on symptoms. Beyond sleep health, our software addresses broader medical issues, demonstrating the convergence of healthcare and technology. In summary, our platform empowers users to proactively manage their health through data-driven insights and accessible virtual consultations, significantly advancing healthcare accessibility and personalized management.

What we learned

Through this project, we've learned invaluable lessons in EEG data analysis, software development, machine learning, healthcare integration, overcoming challenges, and team collaboration. Analyzing EEG data using Convolutional Neural Networks enhanced our understanding of sleep patterns. Building the front end with ReactJS and the back end with Flask, NodeJS, ExpressJS, and MongoDB deepened our expertise in web development and database management. Implementing machine learning algorithms for specialist recommendations broadened our knowledge of predictive modelling. Integrating healthcare features like virtual consultations honed our skills in healthcare technologies. Overcoming challenges such as interpreting research papers and integrating real-time features taught us resilience and problem-solving. Effective team collaboration and coordination were pivotal to project success, emphasizing the significance of teamwork in complex endeavours.

What's next for SleepSync

Next, we're dedicated to enhancing our model's efficiency. We'll optimize hyperparameters, explore data augmentation techniques, and explore advanced feature engineering. Ensemble methods like model averaging will be considered to boost performance further. Additionally, deployment optimizations such as model compression will reduce inference time, ensuring efficiency in production. These efforts aim to elevate the accuracy and reliability of our model, providing users with enhanced insights into their sleep health.

Built With

• cnn
• deep-learning
• express.js
• flask
• graphql
• machine-learning
• mongodb
• node.js
• react
• socket.io
• webrtc