Use cases of computer vision in retail

avatar
Mikołaj Saik
Use cases of computer vision in retail

Retail is changing rapidly in the time of a COVID-19 outbreak. Most of the markets observe that innovation adaptation during the crisis has accelerated a lot. According to McKinsey, organisations that can quickly reimagine their omnichannel approach to create a distinctive customer experience will recover much quicker after the pandemic. CX investments are the key to overcome the downturn without any major problems. Let’s see how CX leaders dealt with the crisis from 2008 in comparison to market followers:

 

McKinsey outlines 5 basic actions that retailers should take to boost the customer experience and adapt to the current economic situation. Two of them, that concern omnichannel innovations and the transformation of the physical stores, are being executed with computer vision technology.

What is computer vision?

According to SAS Insights, computer vision is a field of Artificial intelligence that trains computers to interpret and understand the visual world. Using digital images from cameras and videos and deep learning models, machines can accurately identify and classify objects — and then react to what they ‘see’. Regarding the retail market, the algorithm is supplied with input data in the form of images and videos sent from shop cameras or users' smartphones. On their basis, it can identify objects, people or patterns. Then, the image is analysed and processed in order to generate the data needed to run a business and make strategic and operational decisions, as well as to protect the goods against fraud. The goal of this technological development is an easy one: it aims to automate what the human visual system can do.

Computer vision is spreading in retail

Globally, according to the analysis of BusinessWire, the computer vision market is expected to grow from $ 2.9bn in 2018 to $ 33.5bn by 2025. Regarding the retail industry in specific, the situation does not look worse. According to a RIS report entitled ‘29th Annual Retail Technology Study: Retail Accelerates’ only 3% of retailers have already implemented computer vision technology, however, 40% are going to implement it within the next two years. It is worth paying attention to market leaders such as Walmart or Amazon (which has 27 autonomous stores scattered across the state), or Chinese retail giants who introduced a wide range of solutions from this technological stack long ago. Indeed, shelf scanning robots or overhead cameras that recognise customers and their behaviour are nothing new to those networks.


Use cases in retail 

The main use of computer vision in retail includes: 

  • Stock visibility - the awareness of what is basically happening at the store. The camera system coupled with technology is able to see all kinds of fraud attempts and identify customers who behave suspiciously. The result is simple - minimising theft losses, and thus higher ROI over the long term. It is also worth mentioning the stock visibility in the context of stock replenishment. The camera system is able to record deficiencies in specific batches of goods and inform the store "live" about it; as a result, the supplies are added immediately.

  • Cashierless stores - more than once, we have described the innovation of automatic stores provided by Amazon - Amazon Go. In this case, computer vision technology can be treated as a detection tool that can be described as a desire to buy when the customer picks up the product from the shelf. Numerous cameras detect this action as a purchase and charge it for picking up products in a moment that customer is leaving the store. 

 

  • Marketing - based on the recognition of client behaviours through machine learning, the systems can remember specific patternsof individual clients. Thanks to this (without violating their privacy because those customers are usually anonymous), the retailer is able to build a personalised marketing message based on the possibilities offered by this technology. For instance, if the cameras have registered a customer who browsed a certain batch of products, e.g. TV sets, the system recognises them and processes this information in such a way that after leaving the store they will receive marketing communication regarding this equipment, if they haven’t  already purchased it.

  • Merchandising - Computer vision technology is able to analyse behavioural patterns in the store and on this basis create heat-maps of stores. Thus, the analysis of customer behaviour makes it possible to determine issues such as the best store layout to maximise profits, better product layout, which products should be added or what additional promotions should be considered.

 

Examples of computer vision technology providers

Scandit

New partner of Future Mind - Scandit is a company providing computer vision solutions, integrating barcode scanning, OCR and augmented reality, which are a cost-effective and versatile software-based alternative to dedicated physical scanners. Its technological development is based on scanning barcodes and QRcodes using cameras built into mobile tools. Then, in retail, the view received by the camera can be used in a variety of business aspects. For instance, we are talking about managing warehouses and goods that flow into and out of the store. It allows to quickly identify which goods are in packages and which ones significantly improves the work of employees. Additionally, thanks to AR Scandit technology, it gives access to wider information about a given product - for example about its use or users’ reviews. As part of the use of technology in stores, it is worth mentioning the payments made using scanned codes or the Scan & Go solution are provided by Futureproof Retail, which Scandit supports through its line-free mobile checkout technology.

Syte

Syte uses computer vision to implement visual search technology. Visual search engines use videos or photos from mobile devices (or other devices that provide pictures or movies) for searching such results on the global internet. In the case of retail, it is useful in terms of looking for detailed information about the products from the screen. In this case, you can get more information, other variants, opinions or the exact price list of products merely by using the camera on your device. Additionally, the system remembers searched items and informs about discounts or bargains. In-store solutions such as mirrors are a further step in the development of such a product, which can show products in various arrangements. Obviously, technology is not just about clothes. It can be other products as well - DIY or jewellery. 

 

Eyedo

Computer vision also helps with basic operational tasks related to store management such as shelf management, as well as management and tracking of employees. An excellent tool of this type is provided by one of our partners - Eyedo.

The technology which is available on both web and mobile platforms allows you to scan and analyse shelves to detect product deficiencies or other problems related to non-compliance of product placement. In addition, live algorithms allocate tasks for employees, whereby each problem can be solved quickly and efficiently, and their implementation can be seen from the platform. Eyedo, apart from those functionalities, is also an excellent field management tool and valuable in terms of the needs of large retail stores related to employee and task management. If you are interested in shelf management solutions which are based on computer vision or field management just click here. 

 

Future of computer vision?

In conclusion, it can be seen that technology improvement AI which is behind the computer vision technology is still developing. It is worth emphasising that there are some objects which are not recognised at all. However, algorithms learn intensively and technology imperfections will be improved in the near future. A key to the success of perfect interaction and providing users with relevant and personalised visual messages is the reaction of the algorithm to external conditions. The development of computer vision technology can also solve the problems of people who are impaired or blind. Technology is developing rapidly in retail medicine and autonomous vehicles, and it is just a matter of time for the solution to expand into new markets; even those ones that seem unlikely to be implemented at the moment. If you would like to implement computer vision in your business or if you are interested in our partners’ offer give us a shout. We are here to help you.  

Authors

avatar

Mikołaj Saik

blog comments powered by Disqus
7 min. read
50%

More to discover

avataravataravatar
Karolina Antonowicz, Miłosz Michałowski-Żuk, Mikołaj Saik

What are mobile push notifications and what are the best practices for introducing them?

Mobile push notifications are becoming increasingly more popular and are starting to be considered as one of the more common communication channels, while being equally if not even more functional than the others.
10 min. read
avataravatar
Karolina Antonowicz, Mikołaj Saik

Indoor navigation and the best ways to implement

Everyone of us has probably experienced the situation of leaving a store with empty hands. According to the Retailtouchpoints, nearly 96% of shoppers leave a store without making a purchase at least once in their lifetime.
8 min. read

Is there anything we can do for you?

We use cookies to enhance your experience. Read more about cookies in our privacy policy. Agree