Computer Vision: An image tells more than a thousand data points

A picture tells more than a thousand words. At least for humans. But how much does it tell a computer? Until relatively recently not so much. Instead of seeing cute cat pictures and cute puppies, computer programs saw ones and zeros. But that time is over.

By: Jonathan Hvithamar Rystrøm

6. March 2019

How can you create value with Computer Vision?

Thanks to big progress in both algorithms, data and computing power, computers are now able to see. Today, computers can not only tell the difference between dogs and cats but also whether it is a beagle or a border collie. And even the fastest dog connoisseur can't keep up with the speed and scale that computers bring to the table. AI-driven image analysis - also known as "Computer Vision" - is reaching a golden age. More and more companies have opened their eyes to the enormous potential that Computer Vision can unleash. Here's a quote on three considerations that can help kickstart your imagination.

How can Computer Vision create value from images?

If you cut Computer Vision techniques to the bone, they are about one thing: What's in the picture? Although it almost sounds too simple, there is plenty of value to be gained. It can be anything from finding out if lung tissue contains a cancerous tumor to automatically tagging product images so that they are easier to find on a webshop. Even on very complicated images of e.g. a concert audience, Computer Vision can create order in the chaos. It can help count the number of participants and even find out the gender and age of the people!

Where in the organization can you find images that can create value?

For Computer Vision to create value, it first and foremost requires something to look at. Imagery can be hidden anywhere in the organization. Maybe you have a lot of product images on a webshop? Or pictures of damaged products send by customers? The image material can also easily be live, such as. video from security cameras. That way, one can more easily spot potential threats without having to watch hundreds of hours of video.

Where can Computer Vision help generate data?

Maybe your organization does not have a huge stock of easily accessible and value-creating images. Fortunately, AI can also come to the rescue here. With the help of a camera and a good idea, one can get started creating valuable data. One could, for example, set up a camera in the store that could be used to count customers going in and out - it's much more accurate than regular door counters. Or a camera that keeps an eye on whether all the products on a production line or check if a store lives up to quality standards.

Computer Vision can also open up a true goldmine of data: the Internet! In addition to cute cat videos, the internet contains plenty of options for interesting picture data. But the amount of data can be overwhelming with over 67 million photos uploaded a day… just on Instagram! With AI and Computer Vision as curators, it all becomes more manageable. You can e.g. create a system that automatically collects all Facebook posts where your logo or product is included in the image and analyze the reaction they get - even if it is never mentioned in the text of the post.

As you can read, Computer Vision is a field in rapid development. And it does not have to be that difficult - so what are you waiting for?

Share

Other blog posts you may find interesting

2021-07-20T11:19:41+02:00
Go to Top