[G20] - Elephound

• 

  Phound your stuff

• 

  Eli says Hello to you (Onboarding)

• 

  Eli recognized your stuff (Memorize mode)

• 

  Eli found your controller (Search mode)

Ever forget where you put your stuff? We did too. Meet Eli, your personal AI assistant with memory like an elephant. Show her where you store things and she’ll remember. Never search it, Elephound it!

🎯 Inspiration

• Inspired by the frustration of forgetting where items were placed, e.g., while multitasking.
• We wanted to create a solution that saves time and reduces stress.
• The idea: a personal assistant that helps you easily keep track of your belongings.

🐘 What it does

• Elephound is a personal memory assistant.
• It helps you remember where you store your items by “showing” Eli, the virtual memory elephant, where you place them.
• Available as a Mixed Reality application on Meta Quest 3.
• 📅 Future Plans: Mobile device support is in the works to make Elephound even more accessible.

🛠️ How we built it

• Developed using computer vision and machine learning technologies.
• Designed a user-friendly interface specifically for the Meta Quest 3.
• Created an immersive environment where Eli integrates into your physical space.

🚧 Challenges we ran into

• 🌐 Managing a stable user experience despite poor internet connections and unstable AI services.
• Built robust offline capabilities and fallback mechanisms to tackle these issues.

🏆 Accomplishments that we're proud of

• Achieved high accuracy in location recognition within the Mixed Reality environment.
• Successfully integrated the application into Meta Quest 3 while maintaining an immersive user experience.
• Developed a secure, scalable backend for handling a growing user base.
• 🛡️ Overcame challenges with unstable internet and AI services to ensure a consistent user experience.

📚 What we learned

• Gained insights into the complexities of machine learning and computer vision in a Mixed Reality context.
• Enhanced our understanding of user interaction within immersive environments.
• Benefited greatly from discussions with mentors and advisors, guiding us through technical and UX challenges.
• Learned the importance of creating robust systems to ensure service stability under challenging conditions.

🎯 Target Group

• Diverse Audience: Elephound is designed to be helpful for a wide range of users, from busy professionals to forgetful individuals.
• 📱 Future Potential: As we expand to mobile devices, the solution could also benefit elderly users and those with Alzheimer’s, helping them keep track of important items through their smartphones.

🚀 What's next for Elephound (Upcoming Features and Editions)

1. Voice-Driven Features: Plans to allow users to add items or refine their properties by voice, making the system even more user-friendly.

2. More Voice Commands: Further enhancing the user experience with more intuitive voice controls.

3. Smartphone Mode
   Manage and track your items directly from your smartphone. The Smartphone Mode will seamlessly share inventory data with the MR Quest version, ensuring that your item locations and statuses are synchronized across all your devices. Whether you're using your smartphone on the go or your mixed reality headset at home, you'll always have up-to-date information at your fingertips.

4. Elderphound
   Designed specifically to help seniors keep track of their important belongings. Elderphound can also be used to train people with Alzheimer's or other memory-related conditions to remember the most important things in their home and daily life, providing gentle reminders and structured exercises to reinforce memory retention.

5. DIY Edition
   Thinking about a new DIY project? Ask Eli for help! Eli will create a list of all available and missing items needed for your project. Missing items will be researched online, with Eli providing price estimations and checking availability. Items can then be ordered from stores like OBI or other preferred suppliers. Additionally, there is a B2B edition planned for professional sales reps, supporting their daily work in OBI stores by streamlining inventory management and customer assistance.

6. Friends & Familyphound
   Share item locations with trusted friends and family members. Friends or family members can send inquiries about missing items to their network. Each personal Eli assistant searches for matching items, and if a user chooses to share a search result, the friend or family member is notified immediately.

7. Virtual Garage Sale & Price Estimation
   Easily estimate the value of your used items and list them for sale through a virtual garage sale directly within the app.

8. Insurance Claims
   In case of damage to your belongings, quickly create a detailed report for insurance claims. The report includes item details, estimated value, and any relevant documentation, helping you streamline the claims process.

9. Lend & Remind
   Never lose track of borrowed items again! With Lend & Remind, Eli keeps a detailed record of the items you lend to friends or family, including the date, recipient, and a photo of the item. Eli will remind you when it's time to follow up on a borrowed item, ensuring that nothing slips through the cracks and your belongings are returned on time.

10. Fashion Finder
    Need help deciding what to wear? Ask Eli for a fashion recommendation tailored to your current mood or the event you're attending. Eli knows which clothing is appropriate, offering suggestions from your wardrobe that match the occasion perfectly. You can also combine it with "Friends & Familyphound"—request a matching shirt or accessory from a friend for a party at the weekend, ensuring you have the perfect outfit even if it's not in your own wardrobe.

This updated list now includes the enhanced "Fashion Finder" feature, which can be integrated with "Friends & Familyphound" to borrow items from friends, providing a collaborative and personalized fashion solution.

1. Item segmentation Get a separate image for each item capture by an image for easier identification

Built With

• aikit
• meta
• nextjs
• oculus
• openai
• redis
• typescript
• unity
• vercel
• winner
• wit.ai