AWS Machine Learning Blog

Amazon Rekognition is now available in the Asia Pacific (Seoul) and Asia Pacific (Mumbai) Regions

September 8, 2021: Amazon Elasticsearch Service has been renamed to Amazon OpenSearch Service. See details.

Amazon Rekognition is now available in the Asia Pacific (Seoul) and Asia Pacific (Mumbai) AWS Regions.

Amazon Rekognition is a deep learning-based image and video analysis service that can identify objects, people, text, scenes, and activities, as well as detect any inappropriate content. With Amazon Rekognition, customers can easily integrate image and video analysis into their applications with automatic metadata extraction. This metadata can be used in various ways, such as verifying identities for event entry, reviewing user-generated content, enhancing search capabilities, and automating manual workflows.

In this blog post, we will discuss how customers from verticals such as entertainment, social media, and advertising are using Amazon Rekognition to automate manual workflows and gain new insights. We will also provide sample architectures to demonstrate how Amazon Rekognition can be easily integrated with popular AWS services such as Amazon S3 and AWS Lambda.

Verifying identities

Amazon Rekognition provides fast and accurate face search capabilities that can identify a person in a photo or video by searching against a user-created collection of faces. When a user uploads an image or video, Amazon Rekognition detects the face in the image, matches it against the collection of faces, and returns a similarity score (0-100%) to indicate the most likely matches. This feature is helpful for implementing access control to secure facilities such as office buildings. Other common use cases include tagging large amounts of images and video footage for specific individuals, second-factor authentication for financial services, and other Know Your Customer (KYC) workflows.

K-STAR Group, an entertainment company based in South Korea that provides concert ticketing and payment services, is using Amazon Rekognition for their ‘Face Ticket’ service.

“A recurring pain point at concerts is the long line that attendees wait in to provide proof of their purchased paper ticket and then have it validated at the entrance,” says Hyojin Kim, Chairman at K-STAR Group. “To solve this issue, we developed a ‘Face Ticket’ service using Amazon Rekognition. Now, attendees can voluntarily opt-in to quickly verify their purchase using facial analysis rather than waiting in line to get tickets or scan their paper ticket upon entrance. The concerts we support no longer have lines, and the attendees have enjoyed the convenience and fun experience of using our new ‘Face Ticket’ system. When we were developing this service we compared Rekognition with other local facial analysis services, and we ultimately decided to use Rekognition due to its scalability with S3 and the seamless integration with other AWS services.”

To get started on building your own face search applications, see the following blogs and tutorials:

Reviewing user-generated content

As the amount of user-generated content increases, companies have a new challenge to moderate potentially inappropriate content at scale. With Amazon Rekognition, customers can use the Image Moderation API to automatically detect inappropriate or suggestive content. Customers can also customize their moderation needs around what type of content is appropriate for the culture or demographics of their users. This feature can be helpful for any company that has a large amount of on-demand media that require customized content moderation for their audience. is one of the world’s largest online matchmaking services, helping people around the world meet their life partners through a curated database of verified profiles. Great photos are critical to a positive user experience. To ensure photos are appropriate, has a team dedicated to reviewing user-uploaded photos. Automation promised to streamline the process, but the resource requirements in its previous environment were prohibitive. “Automating photo recognition on traditional infrastructure would have required us to procure, deploy, and manage multiple applications as well as new hardware,” says Ajay Poddar, Vice President of Engineering at adopted Amazon Rekognition using a combination of AWS Lambda serverless compute and custom APIs. “Using Amazon Rekognition, we quickly and affordably automated a highly complex process,” says Poddar. “We expect to reduce the time it takes for a picture to appear on a user’s profile by 95 percent, and we have cut manual work by 50 percent.”

For more details on image moderation workflows, see the following example architecture. You can also use a combination of Amazon Rekognition features including image moderation, facial analysis, and text detection to create your own image review workflows.

Example architecture for image moderation

Enhancing search capabilities

With Amazon Rekognition, you can automatically identify thousands of objects (e.g., table, chair, bike) and scenes (e.g., landscape, city, downtown). When analyzing video, you can also identify specific activities happening in the frame, such as “delivering a package” or “playing soccer.” This feature is powered by the DetectLabels API action, which enables users to automatically extract meaningful metadata.

Open Influence, an end-to-end influencer marketing platform, uses the DetectLabels API action to analyze millions of images in order to improve how their platform recommends influencers. Images are run against DetectLabels to extract metadata, such as objects and scenes, and Amazon Rekognition then returns labels for each image. Next, the associated labels are used to power a visual search feature, which enables users to search for influencers by labels rather than by hashtags or demographics. This feature enables clients to quickly source and identify relevant influencers for critical marketing campaigns. Incorporating Amazon Rekognition into their platform has enabled Open Influence to decrease the influencer approval process time by 40 percent, gain 2.6x more insight per post, and cover 30 percent more posts.

For more details on enhancing search capabilities, see the following example architecture which utilitizes AWS Lambda, Amazon Rekognition, and Amazon OpenSearch Service.

Example architecture for image metadata search


Getting started

You can start building with Amazon Rekognition today using the Amazon Rekognition console and by going to the documentation. For Amazon Rekognition image and video pricing details by Region, please see this page.


About the Authors

Venkatesh Bagaria is a Senior Product Manager for Amazon Rekognition. He focuses on building powerful but easy-to-use deep learning-based image and video analysis services for AWS customers. In his spare time, you’ll find him watching way too many stand-up comedy specials and movies, cooking spicy Indian food, and trying to pretend that he can play the guitar.




Kaiser Larsen is a Product Marketing Manager for AWS artificial intelligence solutions. Outside of work you’ll find him hiking, cooking for family and friends, and eating ice cream whenever there’s an excuse to celebrate.