🤔 Problem Statement
- 55 million people worldwide struggle to engage with their past memories effectively (World Health Organization) and 40% of us will experience some form of memory loss (Alzhiemer's Society of Canada). This widespread struggle with nostalgia emphasizes the critical need for user-friendly solutions. Utilizing modern technology to support reminiscence therapy and enhance cognitive stimulation in this population is essential.
💡 Inspiration
- Alarming statistics from organizations like the Alzheimer's Society of Canada and the World Health Organization motivated us.
- Desire to create a solution to assist individuals experiencing memory loss and dementia.
- Urge to build a machine learning and computer vision project to test our skillsets.
🤖 What it does
- DementiaBuddy offers personalized support for individuals with dementia symptoms.
- Integrates machine learning, computer vision, and natural language processing technologies.
- Facilitates face recognition, memory recording, transcription, summarization, and conversation.
- Helps users stay grounded, recall memories, and manage symptoms effectively.
🧠 How we built it
- Backend developed using Python libraries including OpenCV, TensorFlow, and PyTorch.
- Integration with Supabase for data storage.
- Utilization of Cohere Summarize API for text summarization.
- Frontend built with Next.js, incorporating Voiceflow for chatbot functionality.
🧩 Challenges we ran into
- Limited team size with only two initial members.
- Late addition of two teammates on Saturday.
- Required efficient communication, task prioritization, and adaptability, especially with such unique circumstances for our team.
- Lack of experience in combining all these foreign sponsorship technology, as well as limited frontend and fullstack abilities.
🏆 Accomplishments that we're proud of
- Successful development of a functional prototype within the given timeframe.
- Implementation of key features including face recognition and memory recording.
- Integration of components into a cohesive system.
💻 What we learned
- Enhanced skills in machine learning, computer vision, and natural language processing.
- Improved project management, teamwork, and problem-solving abilities.
- Deepened understanding of dementia care and human-centered design principles.
🚀 What's next for DementiaBuddy
- Refining face recognition algorithm for improved accuracy and scalability.
- Expanding memory recording capabilities.
- Enhancing chatbot's conversational abilities.
- Collaborating with healthcare professionals for validation and tailoring to diverse needs.
📈 Why DementiaBuddy?
Asides from being considered for the Top 3 prizes, we worked really hard so that DementiaBuddy could be considered to win multiple sponsorship awards at this hackathon, including the Best Build with Co:Here, RBC's Retro-Revolution: Bridging Eras with Innovation Prize, Best Use of Auth0, Best Use of StarkNet, & Best .tech Domain Name. Our project stands out because we've successfully integrated multiple cutting-edge technologies to create a user-friendly and accessible platform for those with memory ailments. Here's how we've met each challenge:
💫 Best Build with Co:Here: Dementia Buddy should win the Best Build with Cohere award because it uses Cohere's Summarizing API to make remembering easier for people with memory issues. By summarizing long memories into shorter versions, it helps users connect with their past experiences better. This simple and effective use of Cohere's technology shows how well the project is made and how it focuses on helping users.
💫 RBC's Retro-Revolution - Bridging Eras with Innovation Prize: Dementia Buddy seamlessly combines nostalgia with modern technology, perfectly fitting the criteria of the RBC Bridging Eras prize. By updating the traditional photobook with dynamic video memories, it transforms the reminiscence experience, especially for individuals dealing with dementia and memory issues. Through leveraging advanced digital media tools, Dementia Buddy not only preserves cherished memories but also deepens emotional connections to the past. This innovative approach revitalizes traditional memory preservation methods, offering a valuable resource for stimulating cognitive function and improving overall well-being.
💫 Best Use of Auth0: We succesfully used Auth0's API within our Next.js frontend to help users login and ensure that our web app maintains a personalized experience for users.
💫 Best .tech Domain Name: AMachineLearningProjectToHelpYouTakeATripDownMemoryLane.tech, I can't think of a better domain name. It perfectly describes our project.
Built With
- auth0
- cohere
- css3
- html
- javascript
- nextjs
- pytorch
- tailwind-css
- tensorflow
- voiceflow


Log in or sign up for Devpost to join the conversation.