Microsoft introduces Face Recognition, Machine Learning APIs for Azure

Microsoft’s collection of officially supported Cognitive Services offering has grown by three this week.

The company announced the general availability of the Face and Computer Vision APIs (application programming interfaces) and Content Moderator, enabling developers to add artificial-intelligence (AI) and machine-learning capabilities to their applications. Microsoft Cognitive Services is an Azure-backed collection of over a dozen APIs and services that allow coders to bake image recognition, speech, translation and other advanced functionality into their own applications and services.

The Face API, for example, can detect and analyze people’s faces and organize them in accordance to several attributes, including age and gender. Two years ago, Microsoft gave the public a taste of what its face-recognition technology was capable of while offering developers an early taste of the solutions they can build with it.

During Build 2015, the software giant unveiled How-Old.net, a website that invited visitors to upload their pictures so that it could guess their age and gender. It became a viral hit, attracting over 210,000 images within hours of its public debut as it pumped out guesses that ranged from eerily exact to humorously inaccurate. Fast forward to 2017, and the technology has matured enough to help determine a person’s emotional state, according to the company.

Joining the Face API is another technology that can help developers create software that extracts useful information from pictures.

“Computer Vision API gives you the tools to understand the contents of any image: It creates tags that identify objects, beings like celebrities or actions in an image and crafts coherent sentences to describe it,” Joseph Sirosh, corporate vice president of the Microsoft Data Group, stated in an April 19 announcement.

The API can also now detect 9,000 landmarks, both man-made and natural. It can also spot handwriting, allowing applications to apply optical character recognition (OCR) to handwritten material. This feature currently remains in preview, added Sirosh.

Finally, Content Moderator provides machine-assisted moderation of images and text-based content, along with tools that allow for human review. An obvious use case is to detect objectionable content, but it can also be used to help organizations keep an eye out for personally identifiable information that can run afoul of privacy policies or regulations.

To help organizations get up to speed on AI-enabled enterprise applications, Microsoft also announced three new solution templates for the Cortana Intelligence suite of AI-enhanced big data and analytics services. They include a Demand Forecasting and Price Optimization template that provides intelligent demand forecasting and a Personalized Offers template that helps marketers target relevant offers and promotions at consumers based on their profiles.

A new Quality Assurance template predicts production line failures before they catch businesses by surprise. It also helps reduces waste and uncover the conditions that lead to problems with a product or a business’ operational processes.