Machine Learning (ML) is a branch of Artificial Intelligence (AI) that enables software to do more on its own, such as make predictions and offer suggestions, by analyzing and learning from the data it already has. In short, ML helps make software more proactive in the way it supports your business decisions.
There’s a great deal of interest in AI and ML these days, and at Microsoft we’re increasing the rate at which we’re baking these technologies into our products, including Dynamics NAV.
Over the next few weeks we’ll be posting content that looks at how we’re using ML in Dynamics NAV today and provides inspiration and ideas for ways you can leverage the technology yourself.
Today’s post introduces the Computer Vision API, talks a bit about where we use it in Dynamics NAV, and describes how you can access the service yourself.
Using the Computer Vision API for tagging images
We’ll start with the cloud-based Computer Vision API from Azure Cognitive Services. The service uses advanced algorithms to analyze images and return attributes, called tags, that describe what the image depicts. The tags are accompanied by a confidence level that indicates how sure the service is that the tag is accurate.
Here’s an example of what the service “sees” when it analyzes the following image.
Description and confidence:
“a dog lying on grass near water","confidence":0.89067570050173617.
Tags and confidence:
- dog -- 0,999963402748108
- grass -- 0,999726474285126
- outdoor -- 0,997191727161407
- laying -- 0,911336183547974
The confidence is always between 0 and 1, so the service is very sure about these tags.
Image recognition is useful for a variety of purposes. Accessibility in devices and software is a good example. To get some inspiration, check out Seeing AI. It’s a camera app that uses image recognition to help the visually impaired.
To learn more about how the Computer Vision API, see the documentation for the service.
How do we use the Computer Vision API in Dynamics NAV?
In Dynamics NAV 2018, the Image Analyzer extension uses the Computer Vision API to analyze images that are attached to contact persons and items. For example, this is useful for items because it lets you build up a stockpile of metadata about what you sell and use it to fine-tune search results when someone is browsing your webshop.
If you’re interested in the Image Analyzer extension, you can read more about it here.
Getting access to the Computer Vision API
If you want to try out the Computer Vision API, you can use a trial version for free for 30 days. All you need to use the service are a URL and an API key. They’re available right away, and there are a couple of ways to get them:
- Go to the Try Computer Vision API page, choose Get API Key, and then follow the steps.
- If you have an Azure subscription, you can create a new resource on the Azure portal. Choose Create a resource, search for “Computer Vision,” and then follow the steps.
While you’re at it, consider stopping for a minute to check out the other services that are available.
Note: If you add Computer Vision through your Azure subscription you’ll need to allow a day or so before you can use it. If you need the service immediately, use the trial version while you’re waiting.
In the next post about smart solutions, we’ll look at how to make calls to the Computer Vision API from C/AL code.