Visual Search Adopted by eBay
Written by Nikos Vaggalis   
Wednesday, 08 November 2017

Artificial Intelligence is rapidly taking over all aspects of our lives, from calorie counting, to personal assistance and financial advice. Shopping is no exception. Now, with eBay's new powered-by-AI visual search engine, it's even easier as well as more accurate.

The traditional way of searching for that favorite, but elusive, product which you somewhere stumbled upon, was through performing text-based keyword lookups.

For example, say while checking your Facebook activity, you notice an ad about boots but your attention is instead drawn to the sunglasses that the model is wearing. From their unique shape you infer that they must be Ray-bans. So the next logical step would be to Google something like "Ray-ban sunglasses with blue lenses". This query will return a number of product models. But which one is the specific one I've been looking for? Is it model RB3025, RB4239, or maybe RB4221?


Of course, the vaguer the query the vaguer the resulting product listings are going to be. For example, looking for a "T-shirt with a Chinese symbol stamped on it", with symbol being unidentified as well as not being able to be described by a non Chinese user, would most probably return meaningless results back.

And what about the next logical step after identifying the desired model, "where can I  buy it from?". If you rely on Google's you'll be showered with a hefty number of brick-and-mortar stores in addition to online outlets. 

Fulfilling every shopper's most vivid dream of getting at the exact product of one's desire in a snap, becomes a reality with Ebay's new mobile app's in-app features "Find it on eBay" and  "Image search". Enabled by artificial intelligence and machine learning, they allow shoppers to use pictures instead of words to search in eBay’s enormous product database for the exact or next best match.

"Find it on eBay" does not degrade the user's experience as it is totally integrated into the user's daily workflow. So under the same scenario of browsing through Facebook and noticing the Ray-ban sunglasses, "Sharing" them with the Ebay app activates the "Find it on eBay" functionality. The app then just forwards the image to the cloud where a Convolutional Neural Network processes it to create a custom representation which is compared to the stored product listings inside Ebay's database.

It would be interesting to research which parts of this process could be shifted to the user's device rather than exclusively involving the cloud, as newer chipsets allow AI to both live locally and offline. See Google's Teachable Machine - What it really signifies and AI Linux for more.

But it gets even better. There's no need to stumble upon the desired object by luck or randomness. With the other feature, that of "Image Search", you get unconstrained roaming just by using your phone's camera to take a picture of just about anything to begin a search. You just have to click on the camera button within the Ebay app to take a photo, or use an existing photo of the item you want to buy, from your camera roll. Ebay will then show you listings of exact matches or the ones visually similar to what you want to buy.

Both features are integrated into the app and apart from making it dead easy for the user to identify desired objects, they also give a definitive answer to the "where to buy it from" question, directing the potential customers through Ebay's gates. This vendor lock-in raises yet another  question of whether Google will someday become irrelevant, at least as  far as shopping goes, if other brand names follow suit with their own Computer Vision enabled mobile apps.  

That might not be that far away given that augmented reality apps,  have already adopted by a number of household brands, see Amazon's AR View:


or IKEA's place:

“Find It On eBay” is now available and fully incorporated into Ebay's latest version of its Android app, which raises the question of why just for Android and not IOS? Is this yet another data driven decision? Have Ebay figured out that its user base is mostly Android using?

Whatever the case, “Find It On eBay”'s impact on the way we currently do shopping is going to be no less than profound.

ebay logo

More Information

An Easier Way to Search eBay: Computer Vision with Find It On eBay and Image Search Is Now Live

Related Articles

Google's Teachable Machine - What it really signifies

AI Linux

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on Twitter, Facebook or Linkedin.


Rust Twice As Productive As C++

Google director of engineering, Lars Bergstrom, gave a talk at the recent Rust Nation UK conference and claimed that Rust was twice as productive as C++. Given how good Google is at C++, this is quite [ ... ]

Angular and Wiz To Merge

Two web development frameworks used at Google are merging. One, Angular is open source and widely known, while the other, Wiz, is an internal web framework developed and used by Google for some o [ ... ]

More News

raspberry pi books



or email your comment to:


Last Updated ( Wednesday, 08 November 2017 )