Industry News: Visual search offers faster, easier way to find products


This article has been republished with the permission of Furniture Today USA.

Words: Alex Milstein

Humans process visual information faster than text information, and technology company Think Deeply is on a mission to harness the power of visuals for website search.

It created a visual search engine retailers can put on its website to allow consumers to search for similar products based on a starting image. If a consumer has a picture of a sofa and wants to find other sofas that look like that one, they can simply put the image through a search and instantly find several other sofas of that style, color or design.


Rajesh Kommineni and Rajan Rajbhandari, founders of Think Deeply, have a background in the home furnishings industry and have been working on visual search for a few years.

Companies like Houzz have visual search technology available, but Think Deeply is focusing on ease of use and integration. It wants to make visual search simple, fast and affordable.

“You don’t have to integrate our visual search solution or do a bunch of different things; it’s an easy-to-use widget,” said Kommineni. “We come from a home furnishings background, so we understand the industry. Everyone is struggling right now to figure out what they need to do with social media, mobile websites and the showroom experience. One of the things we make certain of our solution is that it’s working well across all channels. It doesn’t matter where the customer is coming from; if they walk into a store with magazine pictures and say they want a product like the one in the magazine, our solution needs to work.”

The deep learning technology used by Think Deeply understands images the same way human brains understand images, said Kommineni. He used the act of identifying a human face as an example.

First, the eyes identify edges and lines in a typical place. Then another part of the brain starts recognizing that, given this image, it’s looking at an eye or a nose or a mouth. Finally, an even deeper part of the brain puts everything together to understand that it’s a face.

David Debusk, head of business development at Think Deeply, said that, similar to an iPhone, this is a type of solution nobody knew they needed.


“Visual search is a reality,” said Debusk. “Visual recognition will just be another tool to help people search. Some people don’t have the vocabulary to describe certain types of home furnishings, but they can take a picture and put it into the search engine. The other thing is, there’s a lot of global commerce happening, and you find people who speak different languages. Image searching will make the whole process less frustrating for everyone. Frictionless.”

Debusk adds that this is the kind of functionality and “everyday magic” that younger generations are comfortable with. They have a high expectation to be able to have these kinds of tools, and the easier retailers make it to find products and do business with them, the better.

The solution is easy for retailers to use. Retailers can give Think Deeply their product feed, and Think Deeply loads it into their systems and then puts a widget on the retailer’s website. Kommineni said it can also be offered as an add-on.


Think Deeply has begun working with a few small- to medium-sized retailers, mostly in the lighting business, and has had success. Kommineni said that in the near future many retailers will start deploying visual search solutions, and one of the goals of the company is to work with large retailers like Amazon and Google.

Think Deeply plans to roll out a few new features soon. One feature, Shop This Look, will allow users to upload an image of an entire room and shop all of the products in the image. The user can search any image and any part of an image to discover products to buy.

Debusk said he thinks this solution will be widely adopted by retailers and that it’s applicable to more than just retail.


“You always have the early adopters then the mainstream laggards, but we are right on top of this now when people are going to do it,” said Debusk. “I think this is applicable not just to retail, but also in business-to-business. You might have someone using it for customer support to take a picture of a broken product and find out what’s wrong with it. Just like we are finding that voice search has become ubiquitous, we will find the same with visual search.”