By Greg Pritchard, Senior Vice President, International Development, GumGum
There’s a quickly growing field of science called ‘computer vision’ that you may not have heard of, but which has the potential to transform everyday life as we know it.
Computer vision is basically the visual branch of artificial intelligence and involves computers that analyse and understand pictures (image recognition, in other words). Thanks to increasingly available supercomputing power and deep learning neural networks, a process that’s loosely based on how the human brain learns, computers using complex algorithms are taught to see and make sense of what’s going on in images on a massive scale.
The very clever individuals who build this kind of software have come to be more officially known as computer vision scientists, but we at GumGum prefer to use the slightly more user-friendly term, image scientists.
Prevalent in academia for years now, image scientists are starting to have a big impact on the commercial world, too. The computer vision technology market is expected to reach over $33 billion by the end of the decade.
A number of sectors -- healthcare, in particular -- are already benefitting from the technology. There are medical developments like the Dulight device — a small wearable camera that uses image recognition to identify objects such as food, money, and traffic signs– which is empowering the visually-impaired in ways we never thought were possible.
We’re also seeing other health benefits come from image recognition, such as Panasonic developing a refrigerator that identifies foods that have gone bad.
Another industry where image scientists are having a major impact already is digital marketing. For example, the technology has given rise to a whole new form of advertising – the in-image ad.
In-image advertising utilises sophisticated image-recognition software to serve ads over relevant editorial images. For example, GumGum’s technology can be used to detect a picture on a news site of a clean-shaven man and serve an ad within it for a razor product. Or it could identify an image of a woman with very white teeth and then serve an ad for a toothpaste brand.
Image recognition is opening up the opportunity for publishers to monetise the editorial images on their platforms, something they’ve never been able to do before, and at the same time offer premium inventory to marketers that they previously haven’t had access to. However, the opportunities are even bigger than that, and it is all part of a phenomenon called the visual web.
Every day almost two billion images are uploaded across the internet. The rise of picture-sharing sites such as Instagram and Pinterest have obviously played a big part in this, as have the proliferation of smartphone cameras, on which most of us are regularly snapping away.
Many brands are desperate to take advantage of this wealth of user-generated content. Surprisingly, 80% of images relevant to a brand don't have relevant text and therefore can't be tracked by traditional social listening tools. That means marketers are still completely unable to take advantage of the photo-sharing phenomenon at scale.
Computer vision has the potential to change all this with software that can monitor and recognise tens of millions of relevant images across the web in real-time. Imagine the powerful new channel this could open up for connecting brands with audiences.
For this reason startups like GumGum and tech behemoths like Google and IBM are emerging as computer vision pioneers. Facebook’s facial recognition technology, for example, is said to be nearly as accurate as the human brain.
As the visual web grows and images replace words as the most important medium of communication (trust me, it will happen), the importance of image scientists and the field of image science is set to skyrocket. Computer vision, image science, image recognition: These are terms that won’t be obscure for too much longer.