Amazon Rekognition is an image analysis service available in the Amazon AI suite. Through the Amazon Rekognition API, enterprises can enable their applications to detect and analyze scenes, objects, faces and other items within images. Rekognition uses the same machine learning and deep learning technologies as other artificial intelligence (AI) services from Amazon.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
The main features of Rekognition include object and scene detection, facial analysis, facial comparisons, facial recognition and celebrity recognition. One common use for the object and scene detection feature is to search or organize a large image library, while a common use for the facial analysis feature is to gauge customer sentiment and demographics.
To enable applications to detect objects or scenes within an image, Amazon Rekognition uses labels. According to Amazon, the Rekognition service supports thousands of different labels, which are applied to specific objects, scenes or concepts found in an image. Labels span a number of categories, such as People and Events, Animals and Pets, Nature and Outdoors, and Transportation and Vehicles. Users detect labels in an image via the DetectLabelAPI, and can request a new label be made via AWS Customer Support.
To add facial recognition and analysis features to an application, customers can use the SearchFaces API and DetectFaces API, respectively. Customers can use the CompareFaces API to compare facial features and the RecognizeCelebrities API to identify images with specific celebrities. Customers can also submit an image to the Rekognition service as either an object in the Amazon Simple Storage Service (S3) or a byte array. Rekogntion supports JPEG and PNG image formats, and images can be up to 15 MB when passed as an S3 object, or up to 5 MB when passed as a byte array.
Amazon Rekognition availability and pricing
Amazon Rekognition is available in the US-East, US-West, EU and AWS GovCloud regions. Amazon charges for the service based on the number of images processed each month, in addition to any applicable storage costs.
Rekognition is also available in AWS Free Tier. New customers who use the service in the Free Tier can analyze up to 5,000 images per month for the first 12 months, and store up to 1,000 faces at no cost.