Revive

Inspiration

Having worked on AI/ML-driven healthcare projects before, we were naturally drawn to the patient safety track. Our inspiration for ReVive came from a desire to build an impactful AI solution that addresses a real-world gap in the U.S. healthcare system—the lack of accessible stroke rehabilitation. Many patients struggle with recovery due to limited access to rehabilitation centers, so we set out to develop a technology-driven, remote recovery solution that empowers patients and clinicians. By combining AI-powered motion tracking with a seamless mobile and dashboard experience, we aimed to make rehabilitation more engaging, accessible, and effective for stroke survivors.

What it does

ReVive is an AI-powered stroke recovery assistant designed to bridge this gap through a mobile app and clinician dashboard. Using motion tracking, real-time feedback, and personalized recovery plans, ReVive mimics proven rehabilitation techniques—including task-specific practice and goal-oriented exercises—to help patients recover anytime, anywhere. ReVive also extends patients’ access to rehabilitation by providing seamless communication with clinicians and real-time progress tracking. By making rehabilitation more accessible and effective, ReVive empowers patients and healthcare providers to improve stroke recovery outcomes beyond traditional hospital settings.

How we built it

We built ReVive by integrating a variety of cutting-edge technologies to create an intelligent and accessible stroke recovery assistant. For the mobile application, we utilized Flutter, ensuring a smooth and responsive experience across devices. Our design process started with Figma, where we crafted intuitive UI/UX mockups to enhance patient engagement. The core of our motion tracking system leverages Google Teachable Machine and TensorFlow for accurate pose estimation, allowing real-time feedback on exercise performance. On the backend, we used Marblism, ensuring secure and efficient data management, while the clinician-facing dashboard was developed using TypeScript, Remix, and Vite, with Ant Design powering the UI and data visualization components. This seamless integration allows patients to perform their prescribed exercises from home while enabling clinicians to track progress remotely, making stroke rehabilitation more accessible and effective.

Challenges we ran into

During the development of ReVive, we faced several technical and analytical challenges. One major hurdle was integrating the camera feed into the Flutter interface, which was crucial for real-time pose estimation but required troubleshooting compatibility issues. We also encountered difficulties with package and dependency installations, which slowed down initial development. On the data side, properly referencing and visualizing patient data posed challenges, requiring careful structuring to ensure accurate representation in the clinician dashboard. Additionally, identifying key data metrics that would be most valuable to primary care physicians for assessing stroke recovery required in-depth research and iteration. Overcoming these challenges helped us refine our approach and build a more effective solution.

Accomplishments that we're proud of

Most of our group members are first-time or beginner hackers, and we are proud of how quickly we settled on an idea and began researching the complications patients and caregivers face during stroke recovery. We were quick to find ways to leverage AI and tools/technologies to realize our vision. We are equally, if not more, proud of our ability to adapt to changing conditions and the varying skills each member brought to the project.

What we learned

Throughout the development of ReVive, we gained valuable insights into both technical tools and the healthcare industry. We learned how to use Figma to prototype and iterate on our ideas, refining our UI/UX to improve accessibility for stroke recovery patients. Working in a fast-paced hackathon environment, we explored new technologies that streamlined development, allowing us to build an MVP efficiently. Beyond the technical aspects, we deepened our understanding of healthcare challenges, particularly the accessibility gaps in rehabilitation. Clinicians are faced with an overwhelming amount of medical information, it’s hard to discern what is important for a treatment. We learned critical metrics and tests such as the ABC-6 and TUG tests are some of the most accessible and validated methods of providing accurate benchmarks for a patient’s progress through stroke recovery. Primary care physicians use these metrics to prescribe personalized sets of stroke rehab exercises and allow the patient to recover most effectively. Most importantly, this experience reinforced the importance of teamwork—effective collaboration was crucial in overcoming obstacles, managing time constraints, and delivering a functional product under pressure.

What's next for Revive

Currently, ReVive is in its MVP stage, demonstrating the feasibility of our vision. While we had an app development pipeline, limited computational resources prevented us from rendering our model into a fully functional Android or iOS app. Despite these challenges, we successfully built a proof-of-concept, with a fully functional web dashboard and a working AI model, including integration components for iOS. Moving forward, our biggest priority is to launch a fully operational mobile application that aligns with our design framework, ensuring stroke survivors have seamless access to personalized rehabilitation anytime, anywhere.

Built With

• antdesign
• figma
• firebase
• flutter
• google-ml-kit
• google-teachable-machine
• remix
• tensoreflowlite
• tensorflow
• typescript