24/7 Customer Service (800) 927-7671

Shazam for furniture has an eye for style

by • August 18, 2016 • No Comments

Say you are flicking through a lifestyle magazine, or looking at a photo of someone’s living room, and you spot a lamp or a chair that may go excellent in your own home. But how do you find that precise one? A new system, created by researchers at Cornell University, can run photos of furniture through a deep learning neural network and select the developer, version and where to buy it.

Googling a vague description of a piece of furniture, like “red leather chair”, is many likely to return a slew of unwanted results, so the Cornell researchers created a neural network in the vein of Shazam, the app that listens to a snippet of music and in seconds can bring up the song’s title, artist, album and other details.

The system created at Cornell scans photos of objects, and compares their shape, color and showcases to a database of “iconic images” taken of developer catalogs and websites. From there, it returns the many match, along with details on who makes it and where it is actually on the market.

“It appears a lot of folks want to buy things they see in someone else’s home or in a photo, but they don’t understand where to look,” says Sean Bell, one of the paper’s authors.

Online communities like Houzz, Pinterest and LikeThatDecor allow users to share information on products, which include where to buy them, but the researchers wanted to automatize
and streamline that system into a service.

Images of household scenes, with boxes drawn around items of interest, were crowdsourced, and fed into a neural network along with their iconic images to train the system to recognize objects. Rather than waiting for it to trawl through the entire database, the range of matches is narrowed down by initially broadly analyzing the image for its edges and lines, to weed out the many obviously wrong answers. So it looks at additional specific parts and shapes, preceding searching for entire objects in a now much more compact list.

The researchers have created a startup company, Grokstyle, to get the service up and running on a subscription basis for retailers and designers, and in next, much like systems may be created for other areas like clothing.

The paper was published in ACM Transactions on Graphics.

Source: Cornell University

Latest posts

by admin • March 5, 2017