In previous years, marketers were arguably skeptical about the potential performance and benefits that artificial intelligence (AI) could actually offer. The idea of machines being able to analyse behaviour, imagery and other types of data to provide services like personalised advertising was considered futuristic.
However today, more and more companies are betting on the power of AI to underpin their growth and are increasingly showing amazing results for customer recommendations, content curation, personalised news, design and user experience (UX).
What is computer vision?
At a high level, computer vision is an area of AI specialising in the analysis of imagery and video of a wide and diverse nature, functioning in a way similar to how humans use sight and thought to make sense of what they see around them.
The idea is not new, having been around since the 60s, when it was proposed as a way to give robots the capability to interact with the world. Since then, advances in technology have enabled its practical use to grow exponentially to a point where we can now find a myriad of image recognition examples in the tools we use in daily life, helping us to automate small but useful tasks and processes.
A simple example from the world of photography shows us how quickly this technology has brought advanced abilities into the hands of the average user. Starting with DSLR cameras, facial recognition first offered the benefits of faster lens focus. Things have quickly advanced from there, to the point where new mobile cameras like those in the Google Pixel or Huawei Pro 20 can analyse the content of the picture in detail before blurring aesthetically unimportant regions or adjusting brightness to make the image look more true to life.
All interesting features, but how can marketers take advantage of the power of this technology? Well, let’s have a look at some examples.
Taking a familiar use case from the world of ecommerce, a straightforward application can be used to find images of similar content. As customers, when we consider buying a specific piece of clothing, AI is already in use to help find those images of similar items to the one we first sought, with the possibility for the brand to present a change in color, material or texture based on our user preferences or previous purchases, allowing for better personalisation.
Programmatic AI can already even take over the tedious task of testing imagery and messaging targeted to customers to continue that optimised experience across other marketing channels.
A more recent use of computer vision in digital marketing is facial emotion recognition. The idea is to use an emotion detector on the faces of customers when they are viewing different items displayed at a store. This new and additional layer of information helps companies to target specific products to the right customer. In addition, by analysing other factors like age and gender it can be used to define new product categories and groupings around the customers that show interest in them.
The same powerful technology is also available in static image recognition, allowing marketers to flip that example around and instead quickly and easily find content containing people or scenes with similar attributes for use in targeted promotional materials.
Until now, most of these computer vision skills were too slow to be used in real time applications and not generally usable or accurate for that matter. Nowadays, on the other hand, the introduction of deep learning has pushed the speed and accuracy of certain applications to human-like levels.
We can think of all of these example applications in the same light. They’re all refining the search sphere. That means they use image recognition to help marketers significantly reduce the work involved in finding and delivering the imagery, assets or materials needed for the task at hand – driving customer engagement.
How we see it
StoryStream is a pioneer in the application of computer vision technology in the area of automotive marketing. By using custom AI to get a better understanding of automotive images, we can provide insightful information and task automation to our customers.
For example, our proprietary AI technology is able to analyse millions of images from different channels like Instagram, Twitter and third party sites alongside a brand’s own content to extract assets containing a specific brand and model of car as well as the story behind the picture.
For instance, we’re able to identify the emotions, age and gender of the people in the image along with the weather, lighting, color scheme and how those people are interacting with the car.
Our AI skills even enable brands to automatically identify and tag the specific sub-version of the car model, meaning we can then easily apply deep, valuable metadata like emissions and fuel consumption details for marketers to work with.
All of these leading skills are our way of refining the content search sphere for automotive marketers, providing them with an unlimited resource of potential stories to help increase engagement and drive car buyers through their purchase journey.
Summary
In the near future, AI image recognition will be an increasingly bigger part of our lives, allowing for amazing new applications even as consumers.
In terms of security, we’ll be able to authenticate, identify ourselves and even pay using only our faces.
In the wider automotive industry, incredible efforts are going into self driving cars, which requires enormous real-time analysis of road scenes and content to select the best action in real time.
For now though, today’s marketer can already benefit from the significant opportunities AI image recognition has to offer in taking advantage of its automation and time saving applications.
