Amazon Rekognition
Amazon Rekognition is a cloud-based computer vision service that makes it easy to add image and video analysis to your applications, providing capabilities such as object and scene detection, facial analysis, face comparison, celebrity recognition, text detection, content moderation, custom labels, face liveness detection, and streaming video analysis using deep learning technology.
APIs
Amazon Rekognition
Amazon Rekognition provides image and video analysis APIs for label detection, facial analysis, face comparison, celebrity recognition, text detection, content moderation, custo...
Capabilities
Amazon Rekognition Computer Vision Workflows
Unified computer vision workflows combining image analysis, facial recognition, content moderation, and identity verification capabilities for developers and security teams.
Run with NaftikoFeatures
Detect thousands of objects, scenes, and concepts in images and videos with high confidence scores using deep learning.
Detect and analyze faces with attributes including age range, emotions, gender, and facial landmarks.
Compare faces across images to determine if they are the same person with a similarity score.
Create searchable face collections to index and search millions of faces in near real-time.
Identify thousands of celebrities in images and videos across categories like sports, entertainment, and politics.
Detect and extract printed and handwritten text from images and videos in multiple languages.
Detect explicit, inappropriate, or violent content in images and videos for automated content moderation.
Build and train custom image classifiers using your own labeled images for domain-specific object detection.
Detect personal protective equipment such as face covers, hand covers, and head covers on persons in images.
Verify that a user is physically present during identity verification to prevent spoofing attacks.
Track and follow identified people across frames in stored video footage.
Identify technical cues and segments such as black frames, end credits, and color bars in video content.
Analyze live streaming video in real-time using Amazon Kinesis Video Streams integration.
Evaluate image quality attributes including sharpness, brightness, contrast, and dominant colors.
Use Cases
Verify user identities by comparing selfies to ID documents or previously stored face images for onboarding and authentication.
Automatically moderate user-generated content on platforms to detect and filter explicit or inappropriate imagery.
Build searchable image and video archives by automatically tagging media with detected labels, faces, and text.
Monitor camera feeds to detect whether workers are wearing required personal protective equipment in industrial settings.
Prevent identity fraud during digital onboarding by using face liveness detection to confirm real users.
Analyze in-store camera feeds to track customer behavior, measure foot traffic, and optimize product placement.
Search video archives for persons of interest by comparing faces against a known collection.
Automatically tag celebrities in photos and videos for media companies to improve content discoverability.
Train custom classifiers to detect proprietary products, logos, brand assets, or industry-specific objects.
Integrations
Process images and videos stored in Amazon S3 buckets directly without downloading content.
Analyze live video streams in real-time by integrating with Kinesis Video Streams for streaming analysis.
Trigger serverless analysis pipelines by invoking Rekognition from Lambda functions in event-driven architectures.
Receive asynchronous notifications when stored video analysis jobs complete via Amazon SNS.
Control access to Rekognition APIs using AWS Identity and Access Management policies and roles.
Monitor Rekognition API usage, latency, and error metrics through CloudWatch dashboards and alarms.
Orchestrate multi-step computer vision workflows using Step Functions to chain Rekognition operations.
Route low-confidence Rekognition predictions to human reviewers using Amazon Augmented AI for quality control.