Good news for Amazon Web Services customers! Amazon is giving its AWS customers access to some serious Artificial Intelligence capabilities it has developed and mustered over the years. The company has announced a string of new, intelligent products including Rekognition, Polly and Lex, that can now be deployed by its AWS customers.
An announcement to the effect was made by Amazon Web Services CEO Andy Jassy at the re:Invent conference which took place in Las Vegas on Wednesday. The timing and nature of the announcement seem to have been aimed at Google and Microsoft Azure — both of whom have been attempting to wrest market share away from Amazon.
Both the companies have been deploying machine learning and AI to create products that promise to make their cloud based services all the more useful. Amazon made use of re:Invent to remind everyone that Amazon is not without some serious AI capabilities of its own. We will be taking a look at each of the products in a series of posts. Here is the first one.
Okay, so first up is the Amazon Rekognition. As the name implies, it is basically an image recognition and analysis service. Just like Google and Facebook’s systems, it is also capable of recognizing objects, scenes and even faces from a submitted image. Amazon says that the tool can be used for real-time analysis, and is so powerful that it can take millions of images in batch form.
Powered by deep learning and built by our Computer Vision team over the course of many years, this fully-managed service already analyzes billions of images daily. It has been trained on thousands of objects and scenes, and is now available for you to use in your own applications. You can use the Rekognition Demos to put the service through its paces before dive in and start writing code that uses the Rekognition API.
The system is surprisingly accurate. On being shown an image containing a dog, it was able to successfully classify the image as an animal, a dog, a pet, and as a golden retriever — even though it was unaware of an relationship between the three.
It did similarly well with images containing human beings. Check this out:
A system this accurate can certainly come in handy in a wide variety of situations.
If you have a large collection of photos, you can tag and index them using Amazon Rekognition. Because Rekognition is a service, you can process millions of photos per day without having to worry about setting up, running, or scaling any infrastructure. You can implement visual search, tag-based browsing, and all sorts of interactive discovery models.
Rekognition can also be used very successfully in situations involving authentication and security. And as Amazon suggested, there are entirely out of the box applications as well, such as building “smart” marketing billboards that collect demographic data about viewers.