AWS Rekognition

During yesterday's #AWSInnovate event, inspired by some of the AI/ML talks I hacked together a quick mash-up to perform image analysis using Amazon's Rekognition API. The demo below is a simple web page which uses pure JavaScript to send an image to an API which analyses the image and sends back the results in JSON format. This page is hosted in an S3 bucket, but the HTML file also works on my Desktop or on any machine with an internet connection and web browser.

As you can see, AWS Rekognition delivers powerful artificial intelligence and machine capabilities, conveniently packaged in a service that makes it extremely easy to use, requiring only basic development/engineering skills -- no knowledge of AI or ML is required. AWS Rekognition uses highly-trained machine learning models to analyse any image, photo or video, and return detailed and accurate inferences/guesses about the contents, including face detection and recognition (mood, age, gender, hats/apparel, position in the photo, etc) and scene and object detection and recognition (ocean, trees, pen, mobile phone, parked cars, skateboard, bike, nebula, planet, etc) -- with a quantified confidence level. AWS also offers a similar service, Textract,for extracting (OCR) and analysing text from documents. It's reminiscent of a child that has learned to recognise some of the things it sees. I've observed similar technologies being used by consumer cloud storage platforms such as OneDrive, and website/CMS platforms such as Wix, to automatically tag uploaded images with potentially-useful metadata.

The potential untapped applications for such rich data obtained from photos and videos are beyond exciting. More on that later, so watch this space...

AI/ML platforms evolve and get closer each year to human capabilities, and will eventually surpass human capabilities -- ultimately enabling AI and Quantum AI (QAI) to pass the Turin Test.

It is inevitable that any species with sufficient technological capabilities will eventually be surpassed and superseded by machines. The evolution from organic to artificial life forms is a natural process that every civilisation must accept.

19 July 2019

Share the love:

Comments: 0

Add Comment


  • Twitter
  • LinkedIn
  • Facebook
  • Windows Live / Messenger
  • Xbox Live
  • RSS
  • Email