HackIreland-Group35-EVA

• 

  main screen

Inspiration

We were struck by the loneliness and boredom many elderly individuals face, especially those confined to bed in nursing homes. We wanted to create something that could bring joy, spark memories, and encourage meaningful interactions with loved ones. That desire to brighten someone’s day became the heartbeat of our project.

What it does

EVA is a voice-activated companion app that runs on a tablet. It gathers personal details like birth year, hometown, and interests, then uses that info to generate nostalgic music, local news, and photos. An on-screen avatar interacts with the user, offering conversation, playing tunes, or showing images. Family members can easily upload photos and updates, making EVA a friendly bridge between generations.

How we built it

We combined several technologies:

• Frontend: React (or React Native) for a user-friendly interface.
• Backend: Flask for handling API requests.
• Speech Recognition: Whisper API or Vosk for converting voice to text.
• Emotion Detection: Computer vision models to read facial cues and voice tone.
• Conversational AI: GPT-4 API for generating natural, empathetic responses.
• Multimedia: Spotify/YouTube APIs for music and Google Images/Unsplash for nostalgic visuals.
• Data Storage: Supabase or Firebase to securely store user profiles and media.

Challenges we ran into

• Speech and Emotion: Integrating real-time voice commands and emotional cues required juggling multiple APIs and ensuring they worked smoothly together.
• Accessibility: Designing a truly hands-free interface for users with dexterity or mobility issues demanded a careful user experience approach.
• Personalization: Generating the right nostalgic content meant fine-tuning our AI prompts and ensuring the data we pulled was both accurate and meaningful.

Accomplishments that we're proud of

• Building a functional prototype that recognizes user emotions and tailors responses accordingly.
• Creating a simple, intuitive user interface that elderly individuals can use without frustration.
• Successfully integrating family uploads so loved ones can share photos and messages in real time.

What we learned

• Empathy is key: the most impactful features emerged when we focused on the real needs and daily struggles of nursing home residents.
• Collaboration is powerful: blending multiple technologies under one umbrella showed us how cross-functional teamwork can lead to creative breakthroughs.
• Small details matter: from voice activation to nostalgic imagery, every piece plays a role in making the experience welcoming and memorable.

What’s next for HackIreland-Group35-EVA

We plan to:

1. Enhance Emotion Detection: Refine our model to better recognize subtler emotions.
2. Expand Content Options: Offer more varied activities, like interactive games or guided reminiscence sessions.
3. Streamline Deployment: Simplify setup so nursing homes can roll out EVA quickly without heavy tech support.
4. Research & Validation: Conduct user testing in real-world nursing home environments to gather feedback and improve the experience.

With these next steps, we aim to make EVA a trusted, long-term companion that truly brightens the lives of its users.

Built With

• empathetic-dialogue-speech-recognition:-whisper/vosk-for-hands-free-interaction-emotion-detection:-computer-vision-models-to-interpret-facial-expressions-and-voice-tone-data-storage:-supabase/firebase-for-storing-user-profiles
• frontend:-react/react-native-for-a-friendly
• intuitive-interface-backend:-flask-for-handling-api-requests-conversational-ai:-gpt-4-api-for-natural
• media