The transformation of Web 2.0 facial recognition software service Riya into shopping visual search engine Like.com is a remarkable story, involving some interesting technology. Like.com was just launched a couple of days ago, but the company behind it almost may have become a Google acquisition last year.
Riya started out with a great deal of buzz and excitement behind it. Originally a way to recognize visual features in photos, it began a metamorphosis into a visual search engine this summer. Now, with the November 8th start date of the new visual shopping site, Riya looks like it has something pretty interesting.
There are a number of other recent posts about the new site. Here are a few:
- The MIT Technology Review: Visual Search for Better Online Shopping
- ZDNet’s Between the Lines: Riya launches Like.com visual similarity shopping
- Riya CEO Munjal Shah’s blog, Recognizing Deven: Riya launches Like.com
I ran across four patent applications assigned to OJOS, Inc., parent to Riya.
All four were filed on October 7, 2005, and published on November 9, 2006. The inventors listed for all four are Salih Burak Gokturk, Dragomir Anguelov, Vincent Vanhoucke, Kuang-chih Lee, Diem Vu, Danny Yang, Munjal Shah, and Azhar Khan. The documents also share the same abstract, and the detailed descriptions in the applications are very similar from one to the next.
System and method for enabling search and retrieval from image files based on recognized information
US Patent Application 20060253491
System and method for enabling the use of captured images through recognition
US Patent Application 20060251339
System and method for providing objectified image renderings using recognition information from images
US Patent Application 20060251338
System and method for recognizing objects from images and identifying relevancy amongst images and information
US Patent Application 20060251292
The shared abstract:
An embodiment provides for enabling retrieval of a collection of captured images that form at least a portion of a library of images. For each image in the collection, a captured image may be analyzed to recognize information from image data contained in the captured image, and an index may be generated, where the index data is based on the recognized information. Using the index, functionality such as search and retrieval is enabled. Various recognition techniques, including those that use the face, clothing, apparel, and combinations of characteristics may be utilized. Recognition may be performed on, among other things, persons and text carried on objects.
The patent applications don’t discuss the use of visual recognition technology in a shopping context in much detail, and their filing date of October 7, 2005, places the documents at the point of time when rumors were circulating that Riya might be an acquisition target of Google. This snippet, with its choice of an example search partner, is interesting:
Under one variation, search module 1730 may make a search request outside of the system shown in FIG. 16. For example, the search module 1730 may submit a search request based on the user-input to a third party network search engine (such as GOOGLE). In one embodiment, if the user input is text, then the request is the text submitted. If the user input is image, then text associated with the recognition of that image may be used.
There’s also a detailed section describing how this technology could be used as a social network. See the “My Photos” section of Riya for a look at how some of that is being done.
Here’s a hint from last October of the potential growth of the technology into a use like that found on Like.com:
According to another embodiment, an overlay (such as shown in FIG. 20 and related embodiments) on the images can be shown when the mouse is on the images. When the user presses on the overlay, the page might be directed to the web page of the actual product item, or full search page of the item from the central server. This way, photos are included to add value to the article, as well as, ads are displayed in images, and in a non-disturbing manner to the user.