With artificial intelligence (AI) and machine learning seemingly permeating just about every facet of the online world, it comes as little surprise to learn that eBay has some similar tricks up its proverbial sleeves.

The ecommerce juggernaut has announced two new computer vision features that are coming to its online shopping emporium this fall.

Image Search & Find it on eBay

Computer vision enables machines to “understand” images by processing and analyzing them on a pixel-by-pixel basis, rather than leaning on human-generated categorization data, such as keywords and descriptions.

Using Image Search, eBay said that you will be able to snap a photo of a real-world object and use it to find similar items on eBay. Similarly, the company is also rolling out Find it on eBay, which allows you to tap items on any other website — for example on a blog or on Pinterest — and find them to buy through sharing the URL with eBay.

Needless to say, it won’t always be accurate — if you take a photo of a unique chair in your local cafe, for example, chances are eBay will just find you a bunch of chairs that look roughly like it. But it will add an interesting extra dimension to eBay’s search function — it’s not always easy to find what you’re looking for using descriptive keywords.

“Moments of shopping inspiration can come at any time, whether you’re walking down the street or browsing your social media feed,” said Mohan Patt, vice president of buyer experience at eBay. “At eBay, we’re focused on creating new complementary technology that helps our millions of shoppers easily find the things they love at the best value.”

Image Search will initially be available through the eBay app on both Android and iOS, though Find it on eBay will be limited to Android.

These new features come hot on the heels of a number of moves made by eBay to expand its AI capabilities — last may it bought out machine learning company ExpertMaker, then a few months later it snapped up computer-vision company Corrigon.

Pictures speak a thousand words

Computer vision has emerged as a key focus for companies across the board. Stock photo giant Shutterstock unveiled a new reverse-image search tool last year that allows anyone to search for pictures based on “look and feel,” (e.g. color schemes or shapes) rather than text-based descriptors.

In terms of how this improves search, here’s an example. Under the old system, searching for photos of “dogs” would lead Shutterstock to suggest “similar images” (see bottom of screenshot) that really didn’t resemble the original photo beyond the fact that it had a dog in it.

Above: Old Search

Using computer vision, it was possible to find far more relevant alternatives.

Above: New Search: Visually similar

Back in March, Pinterest officially launched a new feature called Lens, allowing users to snap a photo with their mobile camera to find themes and pins related to the photo. It later expanded its capabilities by rolling out image-based recipe searches too.

As for eBay, the company said that many of the AI and machine learning technologies that power its new image search features are “already embedded across the eBay experience,” designed to make its catalog of more than one-billion items easier to sift through.

For example, eBay already offers a little “similar items” merchandizing strip at the bottom of every item page, which shows other goods a shopper may be interested in. “It’s machine learning and AI at the very simplest level, and we’ve seen a tremendous amount of return on investment on that,” noted Japjit Tulsi, VP of engineering at eBay, in an interview with VentureBeat earlier this month.

By pairing its gargantuan consumer datasets with AI and machine learning, this could go some way toward steering eBay away from its recent tepid financials, and ensure it keeps apace with competitors such as Amazon.

For eBay, you could say AI is ride or die.

“If you’re not doing AI today, don’t expect to be around in a few years,” added Tulsi, “It really is that important for companies to invest in — especially commerce companies.”

Source: Venture Beat