Microsoft is taking automatic picture recognition a step further in an online demo which recognizes and describes images. According to the home page, for CaptionBot, "I can understand the content of any photograph and I'll try to describe it as well as any human."
In reality, the results are hit or miss, though markedly better than other solutions which only recognize single objects. One picture we tried was identified as I am not really confident, but I think it's a man in a red shirt is holding some food and he seems". Clumsy grammar aside, the artificial intelligence was able to recognize and interpret multiple parts of the same image. The picture in question was of me holding a bag of Skittles. This is one of a suite of online demos which Microsoft is currently hosting including others that can guess how old a person looks and if two people are twins.
Microsoft hopes to use the technology to provide automatic alt tags within PowerPoint and Word, making image-filled documents and presentations potentially more accessible. Those improvements are set to be released in Office 365 sometime in 2017. In the meantime, you can submit your own photos to CaptionBot and see what Microsoft thinks of them. Photos can either be uploaded or referenced using a URL of an online picture. Note that according to the website, photos will be retained for research purposes. Thanks to Chancey for the tip.
Source: CaptionBotCategory: News
No one has commented on this post.
You must be logged in to post comments.
Or Forgot username or password?
Register for free
J.J. Meddaugh is an experienced technology writer and computer enthusiast. He is a graduate of Western Michigan University with a major in telecommunications management and a minor in business. When not writing for Blind Bargains, he enjoys travel, playing the keyboard, and meeting new people.
