Rekognition
💡 Definition
Amazon Rekognition is a service that makes it easy to add image and video analysis to your applications. You provide an image or video, and the service can identify objects, people, text, scenes, and activities.
🔑 Key Concepts
- Object Detection: Identifies labels (e.g., "Dog", "Car", "Beach").
- Facial Analysis: Detects faces, emotions (happy, sad), and attributes (glasses, beard).
- Facial Comparison: Compares a face in one image to another (Face ID).
- Content Moderation: Detects inappropriate or offensive content.
- Text in Image: Extracts text from images (e.g., license plates).
⚙️ How it Works
It is a pre-trained AI service. You don't need to train a model. You simply make an API call with the image/video, and AWS returns the analysis results (JSON).
🎯 Use Cases
- User Verification: Face matching for identity verification.
- Media Cataloging: Automatically tagging uploaded photos for search.
- Safety: Moderating user-generated content.
💰 Pricing Model
- Image Analysis: Charged per image processed.
- Video Analysis: Charged per minute of video processed.
📝 Exam Tips (CLF-C02)
- Keyword: "Image analysis", "Video analysis", "Face detection".
- Pre-trained (no ML experience needed).
- Used for content moderation.