Increasing Popularity of Image Based Search Engines

February 2nd, 2007 at 12:00 am
 


At first, the Web contained mostly text-based information. But as technologies progressed so did the kinds of data that could be posted on the Web, from texts to photos, from audios to other kinds of media. By becoming more visually oriented, the Web opened up a whole new era for search engines. This time, search engines needed to be more capable of finding visual contents from images rather than just the usual text-based contents. With millions of photos being uploaded everyday, search engines really need to adapt and develop.

London based Pixsta is just one of the search engine applications that offers image browsing and searching technology. The software searches and automatically extracts the visual contents from the images posted on a website. Just like Pixsta, these kinds of search and browsing technologies are great for e-commerce sites and would be ideal for an online art gallery or a video library.

Pixsta recently launched Elle’s browse&buy and Explore on Net-a-porter.com, both perfect examples of how pixsta is able to "organise large image collections into hyperlinked networks of visual similarity, so that users can browse the network to find images that come close to what they want, and also spark off new searches."

Pixsta’s search technology aims to convert more window shopping browsers into buyers by categorizing and linking the images by setting their shapes, colors, and texture as search parameters. Aside from the two sites above, Pixsta have also set up www.chezimelda.com, a large online shoe store.

 Riya and its shopping site Like.com are direct competitors of Pixsta. Just like Pixsta, search methods are conducted through the use of images. For Like.com, searches are conducted by crawling the webpages of more than 200 online stores and searching the more than two million different items. It seems, however, that Riya is far ahead in technology than Pixsta.

The former already have face recognition technology for searching photos and it was only in late last year that the company entered the consumer product market by launching Like.com. I’ve conducted some searches in Like.com and found the technology adequate. The search engine produce quite good results when colors and texture descriptions were used as search parameters.

Both Like.com and Pixsta were right on one thing, to target the huge consumer market. Because of the image based searches, people who read fashion magazines and watch entertainment channels are the ones that would benefit most from them. If someone sees a bracelet or watch a celebrity wore in a fashion or awarding event, he/she can search what specific watch it is by entering search words the describes what the watch looked like.