AWS Computer Vision Services
On this page (13sections)
Introduction
AWS offers managed computer vision through Amazon Rekognition, which analyzes images and video without requiring you to train a model. It can detect objects, scenes, text, and faces, moderate content, and recognize celebrities. These ready-to-use APIs let developers add vision features quickly.
Definition
AWS computer vision services use pre-trained models to extract information and insights from visual content.
Types
Amazon Rekognition
Image and video analysis service
Amazon Textract
Document text extraction and analysis
Amazon Lookout for Vision
Visual inspection and anomaly detection
Amazon Augmented AI
Human review of AI predictions
Use Cases
- Content moderation
- Document processing
- Face recognition and analysis
- Object and scene detection
- Quality control in manufacturing
Implementation
AWS computer vision services can be accessed via APIs, SDKs, or integrated into applications using AWS Lambda.
In Practice
Rekognition exposes simple API calls for tasks like label detection, face comparison, and content moderation, and Rekognition Custom Labels lets you train on your own images when you need domain-specific recognition. Pricing is per image or per minute of video, with no infrastructure to manage.
Key Points
- High accuracy for common vision tasks
- Real-time and batch processing
- Custom models for specific domains
- Built-in privacy and security
References
- AWS Computer Vision — Overview of AWS computer vision services