🌍 AI-GeoVisionID

Face Recognition of World Leaders using YOLOv8 + ArcFace

App link: https://compvision-ai-geovisionid-facerecognitionsystem-3fzlspcrvy7ayv.streamlit.app/

---

📌 Project Overview

AI-GeoVisionID is an interactive computer vision application that detects and recognizes the faces of world leaders in images and videos. The system combines YOLOv8 for fast and accurate face detection with ArcFace embeddings (via DeepFace) for identity recognition. Unknown faces are handled using a similarity threshold.

Key features:

• Detect faces in images and videos.
• Recognize world leaders using ArcFace embeddings.
• Display identity metadata: Name, Position, Age, Nationality.
• Handle unknown faces with cosine similarity thresholding.
• Streamlit interface for interactive uploads and visualizations.
• Export processed videos with annotated bounding boxes and labels.

---

🛠️ Tech Stack

• YOLOv8 (Ultralytics) → Real-time face detection
• ArcFace (via DeepFace) → Facial embeddings for recognition
• OpenCV → Image and video processing, annotation, video export
• NumPy → Array operations and similarity computation
• Streamlit → Web interface for uploads and visualization
• TensorFlow → Backend for ArcFace embeddings
• Python Standard Libraries → File handling, temporary storage

---

🎯 Usage

1. Clone the repository

git clone https://github.com/yourusername/AI-GeoVisionID.git
cd AI-GeoVisionID

AI-GeoVisionID/
├── app.py                # Streamlit main app
├── yolov8s-face.pt       # YOLOv8 face detection model
├── embeddings/           # Precomputed ArcFace embeddings (.npy)
├── requirements.txt      # Python dependencies
├── detectedfaces/          # Sample images/videos
└── README.md