Brain-mimicking software interprets photos to find clothes and matching items
A new technology, which takes any photo of an outfit and recommends similar items, is poised to revolutionise mobile retail. The FindSimilar software, created by London based Cortexica Vision Systems, was made publicly available ahead of the busy Christmas shopping season in the US and several leading US fashion services and retailers have shown interest in the technology which is being integrated into websites and mobile phone-based apps. Similar services are expected to launch in the UK early in the New Year.
The software mimics the way the brain processes images and finds similarities. A picture of a dress, a blouse or a shirt can be analysed by the software which then delivers similar alternatives. Search results are based on a combination of pattern, style, colour and overall design. This broadens choice and helps shoppers to find items that are more affordable or simply closer to their personal taste.
Online fashion search engine ShopStyle has integrated the software into the new version of its free app whilst ‘StyleThief’, which relies entirely on images for search, has also integrated the software. A shopper can use such apps by simply taking a picture on the camera built in to their smartphone. A quick snap of a shop window mannequin, a magazine picture of an item of clothing, someone in the street or a catwalk model is all that is needed to look for similar items, which are then presented for potential in-app purchase.
Visual recognition specialists Cortexica, based in London, England, have developed the software. The Cortexica software replicates the way that the eye and the brain work together to recognize patterns. This process, which has evolved over millions of years of evolution, is now little more than a smartphone camera click away. Half a million items of clothing, with full stockist details for potential purchasers, already sit on virtual clothes rails on Cortexica’s servers.
Iain McCready, CEO of Cortexica, explains: “Our benchmark for this software was to develop something intelligent and discerning enough to satisfy the Miranda Priestlys of the World. We’re delighted that our technology has been so well received. For a retailer, having an app powered by our software is a bit like putting your own shop assistant into a competitor’s store.”
He adds: “We all recognize that deep feeling of frustration after hunting for an item of clothing that we’ve seen or admired or the experience of finding something and wishing we could find a better or sometimes more affordable alternative. Our software is the answer to these perennial problems.”
The software, which will also enable shoppers to take visual clues from other sources such as wallpaper or color swatches and deliver appropriate results, is adaptable and able to learn over time. It mimics calculations made by the human brain when processing images. By doing so, the software finds visual key points of interest tied to patterns. The FindSimilar technology works with images and videos, opening up an array of opportunities with YouTube videos, Pinterest and Instagram images and many more, since it can turn this sort of content into something that is essentially searchable.
WHAT M-RETAILING THINKS: This sort of technology is a significant development in retail and will be what truly integrates mobile devices into the whole real world and online retail worlds. In essence this is the kind of intelligent visual search that will really help consumers find what they are looking for, particularly in fashion, home furnishing and gadget shopping. To date, online retail is hampered by search being semantic, when often goods are desired visually. The sort of tech combines the two to offer a much richer way for consumers to find what they are looking for.
Equally, it helps retailers turn their visual collateral into something much more useful. It will be interesting to see the technology in use and see how well it captures the imagination of retailers. This is certainly the kind of tech to watch in 2014.