Tech can make us think and, sometimes, laugh - even if not on purpose. Microsoft’s Cognitive Services department have released CaptionBot, a new online experiment. Upload a photo to the site and it generates a caption that describes what the algorithm sees. People are then able to rate how accurately it figured out the image. It learns from the rating, theoretically helping the captions get better. (And for robots to officially start taking over the world.)

The bot contains facial recognition, data categorisation and emotion recognition learning abilities. Supposedly able to recognise people, their emotions and other objects in a scene.

"I can understand the content of any image and I'll try to describe it as well as any human," said CaptionBot. "I use Computer Vision and Natural Language to describe contents of images. I am still learning, so sometimes I get things wrong.”

Try it out for yourself, or have a look at some of CaptionBot’s past success and failures.

Leave your comments and see what others say

We would be happy to hear from you.
Login and leave your comments below. How easy is that?


0 total
Be the first to leave a comment