: Revisiting Visual Search as It Should Be

SMS Text

Riya launched, a visual search engine in the year 2006. Image search in its true spirit should be about the image only. Most image search engines today rely on additional tags and meta data to refine their image searches. Its more like feeling for a color with your eyes closed.

Riya, having accomplished at Facial Recognition, launched the engine as a true visual tool that used the image content along with the meta data to describe the image. A set of 10,000 plus variables are used to define the many details in an image as a digital signature. The signature is also mapped to the keywords that get associated with the image. The end result is a mapping from keywords to images factoring in the actual content in the image.

Carsten from SEJ had reviewed when it was launched. The search is tailored towards shopping, jewelery, clothing and other accessories. While truly global image search is as such absent, the technology does provide a path to image search purely driven by content.

There’s a load of information for entrepreneurs to have from the blog of’s CEO Munjal Shah. Tid-bits from them reveal that company paid $1,00,000 for the domain name and was once in talks with Google.

The potential for such technology is indeed immense and with the current social trends online, image recognition is a happening avenue.

Get the latest news from Search Engine Journal!
We value your privacy! See our policy here.
  • Small Business Marketing

    This is really breakthrough technology and it will be interesting to see how quickly they can improve on the search algorithim.

  • SEO for Yahoo Small Business – Terry Reeves

    I would think the process of image mapping could be made easier if the people who were supplying the images accurately described the image.

    How do you teach a computer to recognize similar images if those images could be named anything?

  • Arun Radhakrishnan

    That would have worked in a perfect world :). But the greatest achievement of technology is to use the most elementary of parameters to refine results.

  • Peter

    You should experiment with the engine a bit. You’ll see that the searches frequently don’t make sense.