blog

Technology | April 12, 2013

Image-Recognition Technology and e-Commerce: The Next Frontier?

A recent article in SmartPlanet.com is titled “Madrid start-up will absolutely change the way you shop.” How many times have you heard that a new digital technology will “absolutely change” something? But even if the article doesn’t turn out to be accurate where a particular start-up is concerned, the technology it describes no doubt will.

The start-up is called Visiza and its consumer-facing technology will be part of a website called ShotnShop.com. We’ll let Visiza’s website explain the idea:

Shot & Shop [shotnshop.com] is a visual search engine for fashion products. Shot & Shop recognizes products directly from images of the real environment to provide customers with a selection of visually similar items. Shot & Shop transforms customer demand for fashion products into sale leads for retailers.

So if you see someone wearing something you must have, take a picture of it, upload it to ShotnShop.com and let it search its cache of retailer-provided images to help you identify the product. Then go on to price-compare, find local outlets, etc.

It’s the picture-is-worth-a-thousand-words idea: If the product has distinct characteristics – the SmartPlanet article suggests a purse covered in bottle caps that contain images of fruit – a picture will convey the key attributes better than a verbal search query will.

We suspect that a quirky item like a fruit-bottle-cap-covered purse would actually be pretty easy to find through a verbal search, given its novelty. Something like a standard coffee cup with a certain hard-to-describe decorative pattern might be a better example of something a visual search engine like ShotnShop would do better.

But if it takes off, SmartPlanet notes, the technology would help participating retailers in additional ways, as they could analyze the resulting data for shopping patterns and preferences and geo-location information.

And Shot & Shop sounds like it would perform its specific task better than, say, Google Images does now. As an experiment, we uploaded a photo of our mostly bone-colored coffee cup to Google’s image search page. The search returned no coffee cups: most images were wide-angle shots of people and items that shared the same color palette, including a swatch of red in one corner corresponding to a red pen in the original image.

When we uploaded a carefully cropped picture of a brown lug-soled boot, we got numerous images of baked goods and one of a sleeping cat whose fur matched the color of the boot.

Thus the opportunity for a start-up like Visiza. And recently the Schawk blog talked to Mark Silva, Senior Vice President of Emerging Platforms at Anthem Worldwide, one of Schawk, Inc.’s creative agencies, who told us how image-recognition technology can revolutionize the in-store experience, including checkout:

For consumer packaged goods, I’ve been seeing some solid image-recognition software, and our own BLUE graphics lifecycle management software has created a database of millions of fully rendered objects of packages. If we uploaded that to a shopping app, with a 5-to-8-megapixel camera you could take a picture of an item at any angle and have a fully recognizable, fully rendered object in-app, and once captured you could sync and redeem coupons, recipes or offers and check yourself out, using your phone as you shop – not at the end.

Visiza’s technology and Silva’s prediction support points often made by Lor Gold, SGK’s Global Chief Creative Director: Often, a product is a package unto itself, with its own distinct look that lots of smart marketers have created; and brands should market “from the shelf out,” with the packaging’s form and function having powerful value when carried through to other marketing communications – and soon via searchable smartphone images.