MIRAGÉ

Transform your style one reflection at a time.

Comment

SIGNUP HERE!

Cool stuff ahead, signup [HERE] using [THIS LINK]!

Inspiration

Choosing an outfit should be effortless, but for many, it’s a daily struggle of indecision. With media overload, shifting fashion trends, and the paralysis of too many choices, getting dressed can feel like a chore. The idea of a smart, playful, AI-powered mirror emerged as a solution to this everyday problem—a tool that transforms a mundane task into a delightful experience while reclaiming valuable time and energy.

How we built it

We built MIRAGÉ using the Qualcomm RB3 GEN2 Vision Kit, an IoT development kit powered by the QCS5430 processor. This platform provides scalable CPU, GPU, AI, and camera capabilities, forming the hardware backbone of our smart mirror.

Core Technologies

  • AI and Machine Learning: The application integrates advanced AI technologies, including OpenAI's GPT-4o for contextual responses, YOLOv11 for real-time person detection, and Llama 3.2 (via Lambda Labs) for enhanced text generation. These tools enable real-time analysis of user appearance and generate witty, personalized feedback.
  • Computer Vision: OpenCV handles person detection and tracking, ensuring accurate and reliable interactions with the user.
  • Real-Time APIs: The OpenWeather API provides up-to-date weather information, enabling weather-aware outfit suggestions.

Features and Capabilities

  1. Person Detection:

    • Real-time person tracking using YOLOv11 with adjustable confidence thresholds and detection regions.
    • Automatic presence tracking with cooldown functionality to ensure efficient processing.

  2. Fashion Analysis:

    • Five unique critic personalities offering tailored feedback:
      • Kind & Child-Friendly: Supportive and encouraging advice.
      • Professional & Balanced: Constructive styling recommendations.
      • Weather-Aware: Contextual fashion advice based on weather conditions.
      • Ultra-Critical Expert: Rigorous and high-standard critique.
      • Savage Roast Master: Playful and dramatic commentary.

  3. Interactive Controls:

    • Switch between personalities, adjust detection settings, and trigger responses using keyboard controls.
    • Space: Trigger next critique.
    • Backspace: Skip current critique.
    • Q: Quit application.

  4. Weather Integration:

    • Real-time weather condition monitoring via OpenWeather API.
    • Temperature-specific and condition-aware fashion advice displayed on the mirror.

  5. Discord Bot Integration:

    • Remote monitoring, fashion critique sharing, and image-based feedback.
    • Dedicated channel for mirror interactions, enabling a connected experience.

Hardware and Software Integration

  • Physical Components: The mirror integrates a camera, display, and 3D-printed mounts, alongside speakers for audio output. The hardware supports lightweight edge computing, making it compatible with devices like the Qualcomm RB3 Vision Kit, Raspberry Pi, or Mac Mini.
  • Software Architecture: Python-based, featuring modular components like prompt management (personality systems), weather services, and Discord integration.

Future Capabilities

  • Augmented Reality: Adding virtual try-ons to allow users to see outfits in real-time.
  • Expanded Personality Modes: Introducing more tailored and nuanced feedback styles.
  • Deeper API Integration: Pulling additional contextual data (e.g., calendar events) to enhance suggestions.

By combining cutting-edge AI, robust hardware, and seamless API integration, MIRAGÉ demonstrates the future potential of interactive IoT devices in everyday life.

Challenges we ran into

The team was familiar with Arduino Microcontrollers and RaspberryPi Dev Boards The Qualcomm Robotic Dev Board RB3 Gen2 was a challenge to initialize the serial port. Boards are not consistent, and power-up behavior odd Qualcomm documentation regarding connectivity No soft reset, hardware rest only

Accomplishments that we're proud of

Tackled an experimental piece of hardware with an amazing team of hackers A serious “Pivot” on the direction and theme of the of the project Multiple devices and platforms integration Leverage 3D printing

What we learned

How to Pivot, and pivot and pivot…. Bring more prototyping tools and materials for the physical model New hardware development platform, slices, and mechanical equipment

What's next for MIRAGÉ

Expanding Horizon As we look to the future of our AI-powered fashion mirror, there are several exciting avenues for expansion that leverage augmented reality (AR) technology, partner branding, and user customization. Here’s a glimpse into what’s on the horizon:

  1. Augmented Reality Integration
  2. Virtual Try-Ons: Users can visualize outfits on themselves in real-time using AR, enhancing the decision-making process and reducing the need for returns.
  3. AR Fashion Shows: Host virtual fashion shows where users can see how outfits move and fit in a dynamic environment, adding a layer of excitement and engagement.

  4. Interactive Styling Sessions: Users can interact with their AI assistant through AR, making styling suggestions more immersive and personalized.

  5. Partner Branding Collaborations

  6. Fashion Brand Partnerships: Collaborate with popular fashion brands to curate exclusive collections that users can access directly through the mirror, creating a unique shopping experience.

  7. Influencer Collaborations: Work with fashion influencers to provide curated looks and styles, allowing users to follow trends set by their favorite style icons.

  8. Local Boutique Features: Highlight local shops and emerging designers, promoting community fashion and providing a platform for new talent.

  9. User Customization Features

  10. Personalized Style Profiles: Allow users to create detailed profiles, including style preferences, body measurements, and occasion-based needs, for highly tailored recommendations.

  11. Mood-Based Outfit Suggestions: Introduce features that suggest outfits based on the user’s mood or specific activities, making dressing a more intuitive experience.

  12. Customizable AI Personalities: Let users choose or customize their AI assistant's personality, making interactions more relatable and enjoyable.

    1. Community Engagement

  13. User Style Sharing: Create a platform where users can share their outfits and styling tips, fostering a sense of community and collaboration.

  14. Feedback Loop: Implement a system where users can give feedback on outfit suggestions, allowing the AI to learn and adapt to individual tastes over time. Conclusion These expansions not only enhance the user experience but also position the fashion mirror as a versatile tool in the evolving landscape of fashion technology. By embracing AR, fostering partnerships, and promoting customization, we can create a truly innovative platform that meets the diverse needs of modern consumers while keeping fashion fun and accessible.

Built With

Share this project:

Updates