How to setup Amazon Rekognition in eMAM?
You must make the following configuration level changes to use Amazon Rekognition as an AI service for indexing video assets in eMAM:
AWS Console Configuration
1. AWS Console: Create your account by signing up with AWS console.
2. IAM (Identity and Access Management): Under AWS Services, from the list of AWS Services, search for ‘IAM’ in the search bar.
2.1. Users: Select Users and click on Add User.
2.1.1. Enter User details and access type based on the type of user:
• User name
• Access type: Programmatic access/ AWS Management Console access
2.1.2. Go to Attach Existing Policies and enable check boxes for AmazonRekognitionFullAccess and AmazonS3ReadOnlyAccess permission for the IAM User.
Click on Next Review button.
2.1.1.3. Review the permissions and click on Create User button.
2.2. Roles: Select Roles and click on Create Role.
2.2.1. Under AWS service, choose Rekognition as service for role use case and click on Next Permission button.
2.2.2. Add the AmazonRekognitionServiceRole permissions policy to the IAM Role and click on Next Review button.
Note: AmazonRekognitionServiceRole allows Rekognition to call the AWS services.
2.2.3. Enter below details to create a role and click on Create Role button
• Role name.
• Role description.
• Trusted entities and Policies are auto-populated.
2.2.4. New Role gets added with below details. Make note of Role ARN displayed in this page which will be used in the Super Admin interface for AI workflow creation.
3. Simple Notification Service: Under AWS services, select Simple Notification Service (SNS).
3.1. Topics: Select Create Topic from SNS dashboard.
3.1.1. To create a new topic, add Topic name and Display name and click on Create Topic button.
3.2. Subscriptions: Now add Subscription to the AI service callback endpoint running in the Gateway.
3.2.1. Select Create Subscription under SNS dashboard.
3.2.2. To create a subscription, enter details and click on Create Subscription button.
• Topic ARN: (will auto populate based on the Topic previously created),
• Protocol: Keep the default value as HTTP
• Endpoint: Enter URL to the eMAM Insight service. The HTTP URL format should be like (http://[Ip_Address]/eMAMGateway/eMAMInsightService/InsightAwsHandler.ashx)
3.2.3. Below message is displayed. “A confirmation message will be sent to the subscription end point. Once the subscription has been confirmed, the endpoint will receive notification from the topic”.
3.2.4. Make sure your subscription is confirmed (activated) in order to receive a response from AWS recognition once the analysis is completed.
3.2.5. eMAM Director Interface: Once the request is sent, a confirmation URL will be displayed in the Insight/ Transcript widget in the Director interface. Click on the bell notification icon in the Insights/Transcript widget to see the confirmation URL. Copy this confirmation URL.
3.2.6. AWS Console: Go back to AWS Console Subscription page. Click on Confirm Subscription button. Make a note of Topic ARN displayed in this page which will be used in Super Admin interface for AI workflow creation.
3.2.7. Paste the confirmation URL (copied from Director interface (Insights/Transcript)) to verify the SNS Topic Subscription. Click on Confirm Subscription button.
AWS Console configurations are complete. Now let’s move to Super Admin Console configs.
Super Admin Configuration
1. Create AI Profile: Login to Super Admin Console using your eMAM credentials and go to AI tab.
1.1. Go to AI tab and select your unit.
1.2. Add AI Profile: Add Amazon Rekognition as AI Profile.
1.2.1. AI Profile Name: Enter AI profile name.
1.2.2. AI Server: Select the server where the AI Gateway services are hosted.
(To configure AI services in the eMAM system, under Server Management tab, define the eMAM Gateway URL hosting AI services for the Server. This URL is required for the AI services to communicate back to eMAM Gateway)
1.2.3. AI Type: Choose Insight as AI type.
1.2.4. AI Service: Select Amazon Rekognition as the AI services displayed.
1.2.5. API URL: Enter N/A
1.2.6. API Key: Enter N/A
1.2.7. Click Save to save the changes and Cancel to discard the changes.
2. Create Workflow: Follow the steps below to create a workflow to trigger the Amazon Rekognition AI engine.
2.1. Go to Workflow tab and select your unit and click on New Workflow button to create a new workflow.
2.2. Enter following fields:
• Enter Workflow name
• Enter Description
• Select Event Type from dropdown list.
• Set Conditions for assets in the S3 cloud. Select Field, Operator, Value.
• Set Actions:
o Action Type: Select AI Indexing and click on Add button
o AI Profiles: Select Amazon Rekognition
o Variables: Automatically AI Profile ID is displayed. Additionally, manually insert 2 more variables parameters JSON: SNSTopicArn and RoleArn from AWS console. Take the values from Step 2.2.4 and 3.2.6 mentioned above.
2.3. Click on Save button to save the workflow.
Related Articles
eMAM 5.4 Release, April 2023
EMAM, Inc. is excited to announce the release of eMAM version 5.4! This new release includes several key features and enhancements to help you manage your media assets more efficiently than ever before. In addition to the key features, eMAM version ...eMAM Cloud Platform - Production (eCP-Production)
eCP-Production Deployment About this Guide This guide is intended for IT infrastructure architects, administrators, and IT professionals who are planning to deploy eMAM system (PAAS) on AWS Cloud. Overview eMAM Cloud Platform - Production ...eMAM 5.3 Release, Oct 2021
EMAM, Inc. is delighted to roll out its latest eMAM version 5.3 enriched with some exciting new additions like eFeeder, packaged asset delivery using ATEME TITAN, stitch & deliver using Elemental MediaConvert, instant asset, improved editing tools ...Release Notes eMAM version 5.4
Release Notes eMAM version 5.4 EMAM, Inc. is delighted to announce the release of eMAM version 5.4. This latest update includes several new features and enhancements that will help our users to be more productive and efficient in their video ...How to add new languages to the language list in the Transcript widget?
The Language tab displays the list of languages supported by the AI vendors. Users can select the language from the drop-down list and automatically the displayed transcribes get translated to the selected language. Please follow the below steps to ...