AWS Rekognition – Identifying and categorizing image content fast

AWS Rekognition – Identifying and categorizing image content fast

 

How does a business approach the monumental task of reviewing and then categorizing image and video files – other than hiring a room full of blurry-eyed humans to tackle the task? The tricky thing about sorting image and video files is that there is not much metadata associated with them that describes what is seen inside the file. Metadata is the key to organizing digital information properly and quickly so that it can be useful and leveraged to advance various business initiatives.

 

Amazon’s AWS has a cost-effective solution. Its Rekognition image analysis service leverages artificial intelligence (AI) deep learning methods to automatically analyze images and videos, identify and label objects, people, text, scenes, and activities, and generate file metadata. Furthermore, it can search and compare content.

 

Since its release in late 2016, it has been churning away, analyzing billions of image and video files each day. Through deep-learning AI methods, the technology begins to remember and recognize objects over time and becomes faster and more precise in its analysis.

 

The easy-to-use API means that Rekognition integrates seamlessly with many other AWS solutions, such as S3, Lamda, and Elasticsearch, to create a robust and seamless workflow. This means the bottlenecks and backlogs created by uncategorized heaps of digital image and video content can be a thing of the past.

 

Virtually every industry can benefit from this type of technology. Examples of business value across industries includes:

 

Direct-to-consumer media

Any type of social media site or video/image sharing app needs some way to monitor content to ensure that inappropriate content is identified and removed quickly. One of Rekognition’s most touted features is its ability to scan content quickly and flag unsafe content.

 

Retail

Functionality called sentiment analysis detects emotions from shoppers at brick-and-mortar shops providing a new avenue for generating customer feedback.

 

News & Media

All media assets can be quickly analyzed and categorized. The searchable asset library means less time spent finding the perfect media to associate to a story.

 

 

“Big Brother” concerns

As with most innovations in the realm of AI, there have been some civil rights concerns voiced about Rekognition's role specifically in use in the law enforcement and government agencies to assist in tracking down criminals, suspects or to aid in customs enforcement.

 

Amazon agrees that this type of technology should be used responsibly and ethically, but in the case of law enforcement, they emphasize that it’s the government’s job to write the regulations on how technologies will be used to ensure ethics and civil liberties are not violated.

 

At Amazon’s November 2018 all-hands meeting, CEO Andrew Jassy addressed these concerns, but wanted to remind employees of the positive value of this technology:

 

You see it in the value people are actually getting even after just about a year, year and a half of the service, where Rekognition is actively been used to help stop human trafficking, to reunite missing kids with parents for educational applications, for security and multi-factor authentication to prevent theft.

 

This discussion aligns to similar concerns voiced from multiple tech leaders that ethics and regulations around technology need to be addressed on a national and world level. There should be a balance, though, between establishing boundaries to ensure ethical use of technologies without restricting the innovation of new technologies that will provide net positive results for business and humanity.

 

For businesses, Rekognition is clearly a leading tool for automating tedious, time consuming tasks, allowing organizations to become more efficient in the way they work.