Rekognition
Definition
AWS Rekognition is an advanced image and video analysis service powered by machine learning, enabling facial recognition, object detection, and more.
Use Cases
- Amazon: Automated identity verification and fraud checks for customer onboarding using document and face analysis — Amazon uses Amazon Rekognition capabilities (such as face analysis and comparison) as part of AWS identity verification workflows, typically combined with secure image capture, liveness checks, and backend decisioning systems. (Faster onboarding and improved fraud detection by automating parts of manual review; reduced operational effort for high-volume verification scenarios.)
- PBS: Improving video discoverability by generating metadata from video content — PBS has been publicly referenced as using AWS machine learning services including Amazon Rekognition to analyze video and extract labels/metadata that can be indexed for search and content management. (Better content search and organization through richer metadata, enabling users and internal teams to find relevant segments more efficiently.)
Provider Equivalents
- AWS: Amazon Rekognition
- Azure: Azure AI Vision
- GCP: Vertex AI Vision (AutoML Vision) and Cloud Vision API
- OCI: OCI Vision
Frequently Asked Questions
- What's the difference between Amazon Rekognition and Amazon SageMaker?
- Amazon Rekognition is a ready-to-use API for common image and video analysis tasks (like detecting labels, faces, text, and unsafe content) without training your own model. Amazon SageMaker is a platform for building, training, and deploying custom machine learning models. Use Rekognition when the built-in capabilities meet your needs; use SageMaker when you need a custom model or full control over training and deployment.
- When should I use Amazon Rekognition?
- Use Amazon Rekognition when you need to automatically understand images or videos without building ML infrastructure yourself. Common cases include content moderation (detecting explicit/violent content), extracting text from images (OCR), organizing media libraries with labels, verifying identity with face comparison, and detecting people or objects in video for security or analytics.
- How much does Amazon Rekognition cost?
- Pricing is usage-based and depends on what you analyze (images vs. video) and which features you call (for example, label detection, face analysis, face comparison, text detection, or content moderation). Costs typically scale with the number of images processed or the duration of video analyzed, and some features have different per-unit rates. For accurate estimates, use the AWS Pricing page for Amazon Rekognition and the AWS Pricing Calculator with your expected volume and feature mix.
Category: ai-ml
Difficulty: intermediate
Related Terms
See Also