Microsoft’s Custom Vision lets you build computer vision AI models in minutes

I love Microsoft’s Cognitive Services. It’s taken something incredibly complicated – artificial intelligence and machine learning – and turned it into something that’s not just easy, but also commoditized.

The latest addition to Microsoft’s Cognitive Service is Custom Vision, which lets you create sophisticated computer vision applications, but with a minimum of effort and time.

Cornelia Carapcea, Senior Project Manager at Microsoft, demonstrated the technology on stage at Microsoft’s Build Conference in Seattle earlier today. Creating your own custom vision API models just requires a sample of training data – as little as a few dozen photographic samples– and Custom Vision does the rest.

Once Microsoft has created the model — which takes just a few minutes — you can access it via a REST API hosted on Microsoft’s servers. Carapcea said that it can be used across identifying food and landmarks, and even in a retail environment.

Models improve automatically. Custom Vision can select images that are most likely to add the most gain to your model, allowing for you to manually tag the image, which then goes on to improve overall accuracy and reliability.

You can check out Custom Vision today.