AI boosts monitoring of in-store marketing
AI can now be used for increasingly complex tasks. It is no longer a big budget expense or restricted to big companies. Industry has been using production-line fault detection based on machine vision for many years, and parking halls use number plates for the automatic identification of customers.
AI has become more common, enabling anyone to try out, say, Google’s AI and test how well it recognises text, people or objects in photographs.
Automated recognition as a tool
The monitoring platform of the SnapShop in-store marketing application is integrated with Google technology and can use Google’s AI solution for image recognition. For example, when monitoring campaign launches in different locations, AI can recognize a key advertising slogan or phrase from an image, as the basis for interpreting the campaign as being underway.
If text recognition cannot be used, AI can be trained using sample images of various in-store environments, or different products. Recognition becomes more reliable the more you train AI and the higher the number of sample images. At its most advanced, AI can be used to recognise how many of your own and competing products are on shelves, thereby providing market share data.
Image categorisation as a task for AI
Categorisation of in-store images is important when monitoring product displays. Categories enable images to become more than just images: added value is created from their context, by knowing which products are pictured and which projects they belong to.
In optimum circumstances, images are categorised when sent. On the other hand, if picture-taking is outsourced, or even crowdsourced, the mass of images must be categorised retrospectively. AI is a huge help in this. The sender needs only to take and send a picture of a campaign display: AI will handle image interpretation as a background function. AI can also be used to tag categorised images, which expands the ways in which image banks can be used.