Google’s new software automatically captions images

Google has open sourced its Show and Tell system which will now be available in TensorFlow machine learning library.

Google recently released its ‘Show and Tell’ software which automatically produces captions that accurately describes images.

‘Show and Tell’ is an algorithm which was developed in 2014 at Google’s Research Labs. It essentially makes use of ‘intelligent’ techniques to understand photographic content.

In other words, it examines a large collection of sample images which have already been captioned previously by humans in order to understand the image better.

How it works

Users are simply required to point ‘Show and Tell’ at any photographic image. The algorithm will then automatically generate a caption in English.

The software also received an update recently which Google describes as ‘much faster to train and produces more detailed and accurate descriptions compared to the original system.’

What’s the use

Software of this nature comes to use especially when there is a large collection of pictures that need to be searched and catalogued quickly.

This runs along the lines of tools such as Adobe Lightroom which employs facial recognition to find pictures of specific people.

Google has open sourced its Show and Tell system which will now be available in TensorFlow machine learning library. The code of the system is available on GitHub.