How Algorithms Are Replacing Text Searches with Pictures

Google famously started Google Images after the massive response to Jennifer Lopez’s green Versace dress, worn to the 2000 Grammy Awards.

Google being Google, they quickly applied algorithms to the behind-the-scenes data available in each image file prcoessed. Soon, those algorithms expanded their scope to account for the visuals themselves. Before long, “search by image” became a reliable way to source original pictures, as well as any images related in both composition and subject.

However, now that the technology is available, what can be done with it? Visual recognition systems are desired by everyone from Hollywood effects studios to the military. In between, though, are a range of industries that can benefit from the technology. With the amount of shopping being done on the internet, connecting what people see to where they can buy it is developing quickly.

One company making a go of it is Slyce ( Their app(s) utilize a variety of approaches for visual search, while attempting to encompass every option available to consumers for identifying a product. For example, QR codes are everywhere, yet have struggled to justify their existence on their own. By incorporating them into a broader image-identification engine, Slyce is establishing a multi-pronged approach to product identification.

The more impressive component of this, of course, is the ability for users to take any image, highlight the element they want identified, and then find a range of results from the massive superstore that is the internet. It could be a picture found in a magazine at the doctor’s office; a screenshot from a television show; or even a friend’s picture on Facebook.

The benefit of this technology for commerce is flexibility. Slyce offers a Neiman Marcus-branded app that focuses on the products available in their own lines. With a larger retailer like this, searches of popular styles can find exact or near matches without compromising on what the consumer wants.Google has already shown what a wide-open system is capable of. Specialization of these systems allow their creators to manage the content in a way that provides the best deals to consumers without concerns over scams and out-of-date results.

Of course, specialization is what the internet does best. It’s hard to imagine a more design/product-focused library of pictures than Pinterest. The news that they’ve incorporated their own visual search is simply another sign that text searches have gained a powerful new friend.

Leave a Reply