Welcome to Your Eyez, an innovative app designed to help visually impaired individuals navigate their world with confidence. By harnessing the power of advanced AI technologies, Your Eyez provides real-time audio feedback, painting a vivid picture of the surroundings.
We achieved real-time object detection with a latency of just ~100ms by optimizing YOLOv7 models. Our team built this solution with a focus on performance, accuracy, and accessibility.
- Next.js for server-side rendering and static site generation.
- TypeScript for strong typing.
- TailwindCSS for utility-first styling.
- Clerk for authentication.
- YOLOv7: Powers real-time object detection, running over 10 inferences per second on-device. This ensures immediate identification and tracking of objects, providing crucial information about obstacles and points of interest.
- GPT-4 Vision: Delivers insightful and contextual scene descriptions, capturing details beyond simple object recognition to enhance user understanding of their environment.
- ONNX Runtime: Utilized for efficient model deployment on WebAssembly, enabling high-performance inference directly in the browser without server dependency.
- Custom
AudioGeneratorclass for managing audio playback priorities, ensuring important information is communicated effectively. - Utilizes a
PriorityQueuefor efficient audio scheduling, allowing for real-time adjustments based on user needs. - Dual-voice system for varied audio feedback, enhancing the user experience with natural-sounding speech generated by Deepgram.
- Real-time Object Detection: Announces objects like people, furniture, and potential hazards, aiding in navigation and safety.
- Contextual Scene Descriptions: Provides detailed narratives about the environment, including layout, atmosphere, and activities.
- Customizable Audio Feedback: Personalize volume, voice, and frequency of audio cues to suit individual preferences.
- Intuitive Interface: Gesture-based controls for seamless interaction, allowing users to easily manage app functionalities.
- Enhanced Independence: Navigate environments with confidence and safety.
- Improved Situational Awareness: Gain a deeper understanding of surroundings, fostering a sense of connection and engagement with the world.
- Greater Accessibility: Participate in activities previously challenging, opening new possibilities for visually impaired individuals.
- Sign In/Up: Authenticate using Clerk for secure access.
- Start the App: Initiate object detection and scene description with a simple tap.
- Audio Feedback: Receive real-time announcements and contextual descriptions of the surroundings.
- Manual Trigger: Swipe up for a detailed scene description on demand.
- Mute Phrases: Swipe down to mute specific audio feedback, tailoring the experience to user preferences.
- Sensor Integration: Incorporate LiDAR or depth cameras for enhanced environmental understanding.
- Offline Capabilities: Enable core features without internet connectivity, ensuring continuous functionality.
- Multilingual Support: Expand language options for a broader audience, enhancing accessibility.
- Open-source Collaboration: Encourage community involvement to enhance the app's capabilities further and foster innovation.
Join us in building a future where everyone can experience the world through Your Eyez.
1. Clone the repository:
git clone https://github.com/abdularif0705/youreyez.git2. Install dependencies:
pnpm install3. Set up environment variables:
Create a .env.local file in the root directory and add the following variables:
OPENAI_API_KEY=your_openai_api_key
DEEPGRAM_API_KEY=your_deepgram_api_key
CLERK_API_KEY=your_clerk_api_key
CLERK_FRONTEND_API=your_clerk_frontend_api
4. Run the development server:
pnpm dev