Empowering the visually impaired to locate and identify everyday objects using voice-guided YOLOv8 object detection.
HorusAI leverages state-of-the-art AI and computer vision to provide real-time auditory feedback for object recognition, improving accessibility and independence.
- 🎤 Voice-Guided Interaction – Uses OpenAI Whisper for speech-to-text processing.
- 🔍 Real-Time Object Detection – Powered by YOLOv8 for accurate identification.
- 🗣 AI-Powered Descriptions – OpenAI API enhances object recognition with contextual information.
- 🏗 User-Friendly Interface – Built with Streamlit for easy deployment and accessibility.
- 📦 Lightweight & Scalable – Cloud integration with MongoDB Atlas for data management.
- 🎮 Audio Feedback – Pygame enables dynamic sound responses.
| Technology | Purpose |
|---|---|
| 🐍 Python | Core programming language |
| 🎨 Streamlit | Web application framework |
| 🤖 OpenAI API | AI-powered descriptions |
| 🗣 OpenAI Whisper | Speech-to-text processing |
| 📦 MongoDB Atlas | Cloud-based database |
| 🎯 YOLOv8 | Real-time object detection |
| 🎵 Pygame | Audio feedback system |
| 📸 OpenCV | Image processing |
📁 HorusAI
│── 📝 README.md – Project documentation
│── 📜 newapp.py – Main application script
│── 🎯 yolov8n.pt – General pre-trained YOLOv8 model
│── 🔬 best.pt – Custom YOLOv8 model trained on glasses dataset
📌 Pretrained Model Dataset:
YOLOv8 Glasses Dataset on Kaggle
- Install dependencies:
pip install -r requirements.txt
- Run app in Streamlit!
streamlit run newapp.py