During the last BUILD that Microsoft organized, we have been able to meet different bots with Artificial Intelligence that perform different functions and in which Microsoft is working to launch to the general public. We have already known Tay and his sad future, a chatbot that for the moment is withdrawn from the general public and is now CaptionBot's turn.
CaptionBot is an Artificial Intelligence bot that tries to recognize the elements of images to describe the image and obtain a textual caption to that photo. The CaptionBot intent is not to capture information but to learn to recognize objects or at least this is what Microsoft says on its website. In this regard, the CaptionBot functionality is very interesting because it can be used in the not too distant future for blind or visually impaired people.
CaptionBot effectively recognizes elements of images
To teach CaptionBot and also work as a tool, Microsoft has released a web where this bot has uploaded and where we can test the CaptionBot functionalities, for which we only have to upload the image and Artificial Intelligence will try to recognize all elements or at least the main elements of the image. Those who have already tried the new CaptionBot website and how this bot works say that recognition is pretty good although with some errors, but in general it is quite accurate.
The truth is that I think CaptionBot is an Artificial Intelligence bot more useful than AI bots that currently exist on the market. At least, with CaptionBot we have no possibility of receiving insults or having them speak on our Twitter accounts supporting nefarious ideologies. In addition, it will also serve as a useful tool to process images and obtain your legend for people who do not have very good vision and need additional information.