Artificial intelligence (AI) unlocks many capabilities that would have been unimaginable a short time ago. Amazon’s cloud-based Rekognition tool, part of the suite of solutions under the Amazon Web Services (AWS) umbrella, analyzes images and video to recognize faces, company logos, human sentiment, and more in real-time.
Rekognition can be integrated into applications to provide add-on services, such as the verification of user identities, to insert ads at the appropriate time into online videos or to send smart alerts for deliveries.
See below to learn all about Amazon Rekognition and where it stands in the AI sector:
Amazon and the AI market
The AI market is estimated by Grand View Research to be $93.5 billion and growing fast at a compound annual growth rate (CAGR) of 38.1%. Fortune Business Insights estimates a market size of $328.34 billion, with a CAGR of 20.1%.
Amazon’s Rekognition software can also be considered to be competing in the more narrow AI software market, which Markets and Markets estimated to be $58.3 billion in 2020, with a CAGR of 39.7%. By 2022, Gartner researchers forecast a market size of $62.5 billion, with a CAGR of 21.3%.
Insider Intelligence estimates of $17.78 billion in revenue for AWS in 2022. Key competitors to Amazon in the general AI market include Advanced Micro Devices, AiCure, ARM Limited, Baidu, Enlitic, Google, IBM, Intel, Lifegraph, and Sensely; however, for image and video detection, there are several competitors for Amazon, such as Clarifai, Clearview.ai, Kili, Microfocus, and Google.
Amazon Rekognition key features
- Face detection and analysis via celebrity detection API, image tagging, indexing faces, face comparison and search, recognizing facial attributes (open eyes, sentiment, etc.), searching faces within a collection, searching faces by image, and similarity thresholds for matching faces
- Detect text, optical character recognition, and skewed and distorted text recognition
- Image labeling to detect activities (playing soccer, package delivery, etc.), detect brands and custom labels, and detect objects (trees, bicycles, etc.), scenes (mountains, etc.)
- Video processing features, such as object detection with timestamps, video segment detection (credits, black frames, etc.), and video streaming event detection (package delivery, pets, etc.) in real-time
- Content moderation (recognition of inappropriate content for automated action)
- Model training for the AI
- Command line controls
- Json export
Amazon Rekognition key benefits
When selecting an AI-enabled image or video analytics solution, customers seek the following key benefits from Amazon Reckognition:
Humans can see a face and often recognize people, expressions, or corporate logos at a glance. Computers have much more difficulty, but AI-empowered tools have made great advancements in closing the gap.
AI-recognition enables capabilities for embedded applications, adds efficiency to security monitoring, and provides decision assistance for robotics and self-driving cars. As the technology becomes more widely used, we expect to see many more applications realized.
Embedded application capabilities
Embedded automated recognition can provide an application with quick or even real-time recognition of faces, celebrities, corporate logos, personally identifying information (PII), and more. Once an application can recognize people, places, and things, then the application can start to perform tasks related to image or video recognition.
For example, cameras feeding video or images into Rekognition can issue smart alerts for deliveries or turn on lights for people entering a room. When embedded into smart phone applications, facial recognition can be used by banking software to recognize the user’s face or by help desk software to recognize a user’s level of frustration.
Reality TV shows, user-created content, and influence advertisements create an enormous number of hours of video footage every day. Tired humans screening content often make mistakes and expose broadcasters and streamers for fees and fines related to unlicensed corporate logos to blur, inappropriate content, or exposed PII.
AI-enabled tools such as Rekognition detect inappropriate content quickly and in a standardized manner. They can also detect key video segments, such as black-screen transitions, to speed up and reduce costs for video ad insertion and content production.
Amazon Rekognition use cases
A leader in commercial electronic security systems, 3xLOGIC manages commercial security systems for a wide variety of customers such as corporations, government agencies, medical facilities, and schools. They needed a solution that provided more understanding than simple motion detectors to help eliminate false alarms for their video monitoring services.
“With over 50,000 active cameras in the field, many without the advanced analytics of newer and more expensive camera models, 3xLOGIC takes on the challenge of false alarms every day,” says Charlie Erickson, CTO, 3xLOGIC Products and Solutions.
“Building, training, testing, and maintaining computer vision models is resource-intensive and has a huge learning curve. With Amazon Rekognition Streaming Video Events, we simply call the API and surface the results to our users. It has been very easy to use and the accuracy is impressive.”
The European watch and jewelry maker, Daniel Wellington, strives to provide customer service as high-end as the products they sell to their customers. However, with a number of unique products, employees in their 20 offices worldwide could not quickly identify products and needed AI support to improve their efficiency.
“We realized that a common point of friction for our customers is returning an item,” says Lezgin Bakircioglu, head of global IT operations, Daniel Wellington.
“That’s why we have started working on creating a seamless process using Amazon Rekognition. This image recognition service allows us to automate our process and identify all our products and print labels that expedite our warehouse process. Rekognition has become pivotal to our returns process. With it, we have been able to process returns 15 times faster and with higher accuracy than we could before, thereby eliminating the friction point and creating a better brand experience for our customers.”
National Football League
The National Football League (NFL) receives thousands of media requests for specific players, teams, or plays and needs to search thousands of media assets for relevant events. As the volume of unstructured data grew exponentially, the NFL knew they needed help from an AI to process the data faster and more efficiently.
“By using the new feature in Amazon Rekognition, Custom Labels, we are able to automatically generate metadata tags tailored to specific use cases for our business and provide searchable facets for our content creation teams,” says Brad Boim, senior director of post production and asset management, NFL Media.
“This significantly improves the speed in which we can search for content, and more importantly, it enables us to automatically tag elements that required manual efforts before. These tools allow our production teams to leverage this data directly and provide enhanced products to our customers across all of our media platforms.”
Amazon Rekognition differentiators
When trying to decide between different providers of AI-enhanced video and image recognition, customers select Amazon’s Rekognition because of the following differentiators.
Brand security: AWS
Few companies can match Amazon’s scale or experience in the cloud. When selecting Rekognition customers obtain the security of Amazon and can expect seamless integrations with other AWS services.
Facial and sentiment recognition expertise
Amazon has been training their facial recognition and sentiment analysis features through constant exposure to feeds from their Whole Foods and Amazon Fresh retail outlets. This extensive training enhances Rekognition for security, surveillance, customer service, and marketing purposes.
AWS offers a free 12 -month tier to use and learn AWS features, such as Rekognition. AWS also offers 500+ free digital courses to help build AWS Cloud skills and capabilities quickly. Organizations can test configurations and optimize processes before they start paying.
User reviews of Amazon Rekognition
|User review site||Rating|
|TrustRadius||9 out of 10|
|G2||4.3 out of 5|
|Capterra||4.5 out of 5|
Amazon Rekognition Pricing
Amazon Rekognition can be obtained as part of the AWS Free Tier, in which organizations can test products on a limited basis for up to 12 months. In this free tier, up to 5,000 images can be analyzed and up to 1,000 face metadata objects can be stored per month.
As with other AWS products, the price depends upon the region in which the service is hosted and the volume. The paid tier starts with image processing costs of $0.001 per image for the first 1 million images analyzed, and face metadata storage costs of $0.00001 per metadata per month.
For Rekognition, the type of API (Group 1 or Group 2) does not affect the price until over 35 million images have been processed. AWS provides a price calculator to assist with pricing more complicated needs, including options such as minutes of video processed, detect text, celebrity API, and PPE detection.
Organizations that need to recognize people, places, and things accurately and quickly need AI-powered solutions that can embed into their technology. Cloud-based solutions such as Amazon’s Rekognition provide results with reasonable pricing that can service applications worldwide.
Amazon’s features and capabilities should push it to the top of any list of contenders. However, the AWS free tier allows for a company to experiment with the tool and establish baseline performance benchmarks before investing fully in any specific solution.