The Artificial Intelligence and Machine Learning technologies that come with Google Maps can help you stand out from your competitors. Would you like to know how? We would like to share some examples with you.
Google Maps is the map benchmark par excellence for hundreds of millions of users around the world.
However, the usefulness of its services is not based solely on all the information that it includes and offers. Its power, the aspect which allows it to renew its status as market leader year after year, resides in the constant updating of these data. This gigantic database is made up, in part, of information provided by individual and business users and is constantly being renewed. But, what is the basis of Google Maps’ active role?
You have surely seen the vehicles that take pictures with their 360-degree cameras in the street; they geo-position the relevant information to update the Street View service. These cameras provide millions of images every day and, as you can imagine, it is impossible to analyse by hand the 80,000 million high-resolution images stored to date. To achieve this, the company launched a project called Google Ground Truth designed to achieve the automated extraction of as much information as possible.
Spectacular accuracy in image recognition
There is an interesting document titled “Attention-based Extraction of Structured Information from Street View Imagery” in which Google engineers describe how they are capable of reading the names and numbers of streets quite accurately (84.2%) using deep neural networks. However, this does not only apply to street name signs. It also applies to shop signs and to any other information that may be of interest. Imagine, for a moment, that a sign does not only include the name of a business; it may include other trademarks too.
Machine Learning is also applied to this technology to protect the privacy of individuals (blurring faces or license plate numbers), as well as artificial intelligence techniques to generate panoramic photographs from partial images. In May 2017, they launched their AutoML project dedicated to image and voice recognition. And the results are proving to be spectacular; to such an extent that it is now available as an AutoML Vision service, which makes it possible to train custom vision models for business use without having to be proficient in Machine Learning. And Google is using it to detect objects and classify large-scale images.
What can these services do?
Well, on that very site Google shows us two interesting examples for retail e-commerce related to the recognition of garments based on their visual characteristics, allowing potential customers to filter their searches based on data that goes beyond the name of the garment. Another interesting example is its contribution to preserving nature by detecting animal populations in wild environments to analyse human impact on their migrations and geographic distribution.
Google is also working in the field of healthcare by helping, for example, pathologists to detect cancer in radiological images, understanding medical conversations to help both physicians and patients, and using automatic learning to address a wide range of problems in genomics.
At Intelligence Partner we trust the disruptive innovation and the tools that Google provides to help our customers take advantage of them and transform their businesses. Do you think this innovation by Google might help your business?