Microsoft launched a public beta of its new Custom Vision service on March 1, enabling developers to build mobile applications that can recognize objects.
Part of the Microsoft Cognitive Services collection of cloud-based artificial intelligence (AI) APIs, the Custom Vision service “makes it easy to build and refine custom classifiers” that can be incorporated into apps for iOS, Android and other devices, Microsoft Cognitive Services Principal Group Program Manager Andy Hickl told eWEEK. “Those applications are then imbued with visual intelligence capabilities that run in real time and don’t require a live internet connection to a supporting cloud service.”
Custom Vision can then be used for a variety of purposes, including building intelligent apps that both broadly distinguish between types of objects and correctly identify objects with improved precision.
“With a couple of clicks, Custom Vision service can be used for a multiplicity of scenarios: retailers can easily create models that can auto-classify images from their catalogs (dresses vs shoes, etc.), social sites can more effectively filter and classify images of specific products, or national parks can detect whether images from cameras include wild animals or not,” explained Joseph Sirosh, corporate vice president of Microsoft’s Cloud AI Platform group, in a March 1 announcement.
The technology reflects Microsoft’s mission to commercialize advanced AI capabilities and bring them into the IT mainstream, Hickl added. Available as a paid public preview, he said Custom Vision is a signal to the company’s enterprise customers that “this is something that is for real, this isn’t vaporware.”
Microsoft also unveiled a major upgrade to its Face API. Developers can now tap into the service to recognize up to 1 million faces, enabling them to create applications that can detect “minute differences in more people, with more variation,” opening up a plethora of large-scale personalization use cases, Hickl said.
Finally, the Bing Entity Search service is now generally available in the Azure Portal, Microsoft’s cloud services hub. The service’s related API enables coders to create applications that can extract knowledge from text searches and integrate that knowledge with other systems to add intelligence to apps and enrich the user experience, Hickl said.
Using an extensive catalog of entities—people, places, films, books and more—Bing Entity Search provides users with a “mini knowledge graph,” of sorts, that can extend an application’s utility, like adding location information to photos that appear in social media posts or automatically embedding information on landmarks, restaurants and other entities mentioned in a messaging app.
Another way Microsoft is courting enterprises interested in rolling out their own AI applications and the AI developer community at large is by assuring users that they are in complete control of their data, Hickl noted.
On Feb. 15, the company announced that its Cognitive Services would share the same customer data terms as other Azure services used by enterprises, meaning they own their data and can manage and delete it as desired. The one exception is Bing Search Services, which uses search queries to improve Microsoft’s search algorithms over time.