3 min read

Riya Readies 2.0 Photo Search

Startup allows users to find digital photos across the Web based on similarities in the images.
The boom in search and photo sharing has prompted a startup focused on facial and text recognition to develop a technology that allows users to find digital photos across the Web based on similarities in the images.

Riya 2.0, scheduled for release between October and November, will let users search the Web the way Google Images allows searches, and also conduct similarity searches. Today, the application lets users organize, share, and host photos on the company's servers.

The new technology based on a technique called "image similarity" relies on hundreds of algorithms to look inside the photograph and break the image into grids to identify objects, said Munjal Shah, Riya's CEO. "The searches take between one and two seconds, a little longer than a typical search, but we are trying to figure it out in advance to cut down the search time," he said. "We think we've solved the problem by doing many of the calculations in advance."

The software examines each face or object and turns it into a huge math equation that has about 2,000 parameters. Shah said. Within the software piece that finds similar faces, for example, there's a section to locate similar noises and chins, and even similar foreheads, by looking at the amount of space between the person's hairline and eyebrows.

Today, Riya allows users to upload photos to its servers where they can train the system to recognize faces. The beta site launched in late March, and approximately 7 million photos were uploaded in the first seven weeks. The technology also recognizes text in images, and provides tagging, photo sharing and public and private albums.

But Riya users wanted more. "Many people really wanted the ability to search the Web for all the pictures of them, or a person that looks similar to this person," Shah said.

Susquehanna Financial Group equity analyst Malindi Davies believes Riya's plans could appeal to marketing and advertising firms. "If you have a more efficient way to tag or recognize objects in photographs that's less human-capital intensive that's a great technology for stock image companies like Getty Images, Jupiterimages, or Corbis Photos," she said.

A slew of companies have developed applications that allow users to search the Web for photographs. Greg Isaacs, director of the eBay Developers Program, said two eBay developers, FilmLoop and Slide, have separately developed a desktop application that let users search for photos of product and inventory for sale on eBay. "I suspect one of the databases they will want to search through is eBay, given our millions upon millions of items," he said.

Davies said there's always room in the industry for another competitor, depending on what's unique about their product. Riya will compete with more established photo-search sites, such as Yahoo! Inc.'s Flickr and Google Inc.'s Image Search.