You see a jacket you love in a picture. Perhaps it appeared on social media, in a movie, or while you were browsing online. You want something with the same look, but not necessarily that exact item.
The original may be too expensive, sold out, unavailable in your country, or no longer in production. It might have the right style but the wrong material, fit, or price. You may simply want to explore your options before deciding what to buy.
The item that caught your eye is often only the starting point, not necessarily the destination.
Finding something that keeps the details you like while changing the ones you don't should be easy. Too often, it's a tedious process.
Jumping from retailer to retailer just to find one item can be annoying and time consuming.
The solution to that is what we're building.
Is the jacket cropped or simply short? Is its shape relaxed, boxy, or oversized? What's that particular collar called? Which details make it look different from hundreds of otherwise similar looking jackets?
You shouldn't have to know exactly how to describe what you're looking for. Upload an image, add any details you already know, and FindCloseMatch will search for visually similar pieces.
- 01Upload an imageStart with a garment you like.
- 02We analyze its appearanceShape, fit, colour and design details all matter.
- 03Similar items are rankedThe closest visual alternatives rise to the top.
- 04Browse the resultsExplore options from multiple retailers.
Where we are today
FindCloseMatch is currently an early-stage prototype. The core visual-search technology is working with a research catalogue, allowing us to develop and evaluate how well the system recognizes similar garments.
We want to be transparent about this stage of development. The current version demonstrates the foundation of the product, but it is not yet the complete experience we're working towards.
Start with inspiration, find the piece that gets the details right for you.
We are working toward retailer integrations that will allow the platform to search live products. As those integrations are added, results will include current prices, availability, size/fit options, material and direct links to participating stores.
For our shoppers, that means a shorter route from inspiration to purchase. For retailers, it offers a way to place relevant products in front of people already searching for a particular look.
Why visual similarity is only the beginning
Finding similar clothing is more complicated than matching colours.
Two jackets might both be black and appear similar at first glance while differing significantly in silhouette, length, neckline, material, fastening, or sleeve shape.
Small design details often determine whether a recommendation feels genuinely relevant. We are continuing to improve how FindCloseMatch recognizes garment type, colour, pattern, silhouette, fit, proportions, material, texture and construction details.
Letting shoppers decide what matters
An image can communicate a great deal, but it cannot tell us everything a shopper wants.
For instance, suppose you upload a picture of a black oversized leather jacket. You may want the same silhouette, but only in leather. You might need a particular size, have a maximum budget of, say, €120, or prefer results available from retailers in your country.
Our long term vision incorporates more than similarity alone.
Fit and material help define the kind of product someone wants, while size and availability determine whether they can actually buy it.
Instead of asking only, “Which products look most like this image?”, FindCloseMatch will soon be able to answer a more useful question: which products look similar, have the characteristics I care about, are available in my size, fit my budget, and can be purchased now?
Building a bridge between inspiration and purchase
Retailer product information is an important part of our vision. A product feed can provide more than a link to a store. It can also supply the attributes needed to make search results useful: materials, sizing, pricing, availability, product categories and other details that cannot always be determined reliably from an image alone.
By combining this information with visual analysis, FindCloseMatch can help shoppers compare suitable alternatives without losing sight of what attracted them to the original item. When someone chooses a result, they will be able to continue directly to the retailer to learn more or complete their purchase.
Where we’re going
Our goal is for FindCloseMatch to become a natural starting point when you know what you want something to look like, but don't yet know which product you want to buy.
We are beginning with visual similarity and building toward a more complete fashion-discovery experience. A solution that understands both the appearance of a garment, while letting you explore alternatives like, say, a different type of collar or simply without any print.
There is still a great deal to develop. Search quality will continue to improve, retailer products will replace the current research catalogue, and additional controls will give shoppers more influence over their results.