tessl/npm-pulumi--aws
A Pulumi package for creating and managing Amazon Web Services (AWS) cloud resources with infrastructure-as-code.
—
Pending
Quality
Pending
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
rekognition.mddocs/services/
Amazon Rekognition
Amazon Rekognition makes it easy to add image and video analysis to your applications using machine learning.
Package
import * as aws from "@pulumi/aws";Overview
Amazon Rekognition provides pre-trained and customizable computer vision capabilities.
Key Features
Image Analysis
- Object and Scene Detection: Identify thousands of objects
- Facial Analysis: Detect faces and analyze attributes
- Face Comparison: Compare faces across images
- Celebrity Recognition: Identify famous people
- Text in Images: Extract text from images (OCR)
- Inappropriate Content Detection: Detect unsafe content
Video Analysis
- Real-Time Streaming: Analyze video streams
- Stored Video: Analyze stored video files
- Person Tracking: Track people through video
- Activity Detection: Identify activities and events
Common Patterns
Face Collection
// Create IAM role for Rekognition
const rekognitionRole = new aws.iam.Role("rekognition-role", {
assumeRolePolicy: JSON.stringify({
Version: "2012-10-17",
Statement: [{
Action: "sts:AssumeRole",
Effect: "Allow",
Principal: {
Service: "rekognition.amazonaws.com",
},
}],
}),
});
new aws.iam.RolePolicy("rekognition-s3", {
role: rekognitionRole.id,
policy: JSON.stringify({
Version: "2012-10-17",
Statement: [{
Effect: "Allow",
Action: [
"s3:GetObject",
"s3:ListBucket",
],
Resource: [
bucket.arn,
pulumi.interpolate`${bucket.arn}/*`,
],
}],
}),
});Lambda Integration for Image Analysis
const imageAnalysisLambda = new aws.lambda.Function("image-analysis", {
runtime: "python3.11",
handler: "index.handler",
role: lambdaRole.arn,
code: new pulumi.asset.AssetArchive({
"index.py": new pulumi.asset.StringAsset(`
import boto3
import json
rekognition = boto3.client('rekognition')
def handler(event, context):
bucket = event['bucket']
key = event['key']
# Detect labels
response = rekognition.detect_labels(
Image={
'S3Object': {
'Bucket': bucket,
'Name': key
}
},
MaxLabels=10,
MinConfidence=90
)
return {
'labels': response['Labels']
}
`),
}),
});
// Grant Rekognition permissions
new aws.iam.RolePolicy("lambda-rekognition", {
role: lambdaRole.id,
policy: JSON.stringify({
Version: "2012-10-17",
Statement: [{
Effect: "Allow",
Action: [
"rekognition:DetectLabels",
"rekognition:DetectFaces",
"rekognition:DetectText",
"rekognition:DetectModerationLabels",
],
Resource: "*",
}],
}),
});S3 Trigger for Automatic Analysis
const s3Trigger = new aws.s3.BucketNotification("image-upload", {
bucket: bucket.id,
lambdaFunctions: [{
lambdaFunctionArn: imageAnalysisLambda.arn,
events: ["s3:ObjectCreated:*"],
filterSuffix: ".jpg",
}],
});
new aws.lambda.Permission("s3-invoke", {
action: "lambda:InvokeFunction",
function: imageAnalysisLambda.arn,
principal: "s3.amazonaws.com",
sourceArn: bucket.arn,
});Video Analysis with SNS Notifications
const videoAnalysisRole = new aws.iam.Role("video-analysis", {
assumeRolePolicy: JSON.stringify({
Version: "2012-10-17",
Statement: [{
Action: "sts:AssumeRole",
Effect: "Allow",
Principal: {
Service: "rekognition.amazonaws.com",
},
}],
}),
});
new aws.iam.RolePolicy("video-sns-publish", {
role: videoAnalysisRole.id,
policy: JSON.stringify({
Version: "2012-10-17",
Statement: [{
Effect: "Allow",
Action: "sns:Publish",
Resource: snsTopic.arn,
}],
}),
});Custom Labels for Custom Object Detection
Train custom models for your specific use case.
const customLabelsRole = new aws.iam.Role("custom-labels", {
assumeRolePolicy: JSON.stringify({
Version: "2012-10-17",
Statement: [{
Action: "sts:AssumeRole",
Effect: "Allow",
Principal: {
Service: "rekognition.amazonaws.com",
},
}],
}),
});Use Cases
- Content Moderation: Filter inappropriate images/videos
- Face Authentication: Verify user identity
- Photo Organization: Auto-tag and organize photos
- Security Surveillance: Analyze security camera feeds
- Retail Analytics: Track customer behavior in stores
- Manufacturing QA: Detect defects in products
- Media Analysis: Tag and search video content
Best Practices
- Optimize Image Quality: Use high-resolution images for better accuracy
- Implement Caching: Cache results for frequently accessed images
- Handle Rate Limits: Implement retry logic with exponential backoff
- Monitor Costs: Track API calls to manage expenses
- Use Custom Labels: Train models for domain-specific objects
Related Services
- S3 - Store images and videos
- Lambda - Process analysis results
- Kinesis Video Streams - Stream video analysis
- SNS - Notifications for video analysis completion
Install with Tessl CLI
npx tessl i tessl/npm-pulumi--aws