Reverse image search engine platform TinEye helps photographers see where else their images have appeared online, similar to Google image search but instead of searching a term to find photos that match, users can enter a photo and see where else it has been used online. Today the company is releasing the first of several APIs designed to leverage the company’s image recognition technology for businesses. The API, MulticolorEngine, helps individuals and companies search their image collections or products by color.
TinEye was developed by Idée Inc., which has developed several other image recognition products including editorial image tracking service PixID, and mobile image recognition service MobileEngine, and which counts iStockPhoto and the Associated Press among its clients. The company launched its reverse image search engine in 2008, and is now launching TinEye Services to help app developers and enterprises build image recognition into their apps. The company is currently using the MulticolorEngine to power their own color search tool, which lets users search 10 million Creative Commons images by color. Using the API, businesses can add color search into their web or mobile platforms, for example letting shoppers search by color on clothing ecommerce sites.
TinEye founder Leila Boujnane said MulticolorEngine is just one in a set of APIs designed to help leverage the company’s image recognition technology. Essentially the technology pulls out colors and creates a fingerprint for each image, allowing the platform to match images across the web, and making them searchable by color.
“They basically provide image recognition services for your own applications, your own enterprise, your own enterprise,” she said in an interview about the new set of APIs. “They actually work directly with your own product images.” As for why they’re tackling color matching specifically, Boujnane said there’s no barrier to entry when searching by color, since even if someone doesn’t speak a given language, they can always search by color.
She said anyone with color in their products or site images could leverage the API, for example an ecommerce site that wants to let users shop products by color, or a publication that wants to make archived issues searchable by color. To implement the API, companies work with TinEye to get set up, and then the engine indexes all available images, automatically indexing new images or products as they’re added to the site. It eliminates the need for companies to manually tag their images with colors, and helps extract color palettes for a product, for example pulling out all the colors in a printed dress. It also means that retailers don’t have to change databases or move product information to another system to add color search.
As for shoppers or consumers interacting with sites powered by the API, they wouldn’t necessarily even know TinEye’s technology is being used. For example the company’s MobileEngine API has been powering Snooth’s mobile wine recognition engine, letting consumers snap a photo of a wine label to find out more information.
“We’re pretty much almost like an invisible image recognition engine behind these type of applications,” she said. “We actually want to drive our clients’ implementations rather than being on the front line.”
Companies can access the API via a variety of monthly pricing plans, which all include a setup fee, and monthly subscription plan. The basic starter account also includes a fee for every search and indexed image, so a small online retailer would pay a $500 one-time setup fee, a $200 monthly ongoing fee, then $0.01 per search, and $0.01 per indexed image, and larger retailers have searches and indexed images included in the monthly subscription cost.
Though Boujnane said color searching is new to many brands, she said they are working with several companies, though they can’t announce who just yet. As wearable augmented reality like Google Glass prepares to hit the mainstream, image recognition is quickly becoming the next trend for companies looking to augment their print or online content. TinEye has been a leader in image recognition since 2008, and their new set of APIs should help close the gap between their consumer-focused reverse search engine, and their enterprise-level image recognition offerings.