Now People Can Search Web by Images

Samara, Russia, March 16, 2008 --(PR.com)-- People search for digital pictures and resources in the Internet increasingly often. Conventional search systems are text based, and require certain details or text descriptions to obtain successful results. To find a web page or a picture, user submits textual request. In this case, the names and keywords for desired objects should be correctly spelled with regard to languages, and the query should be stated clearly and carefully. Search result depends on the quality of the textual request, the quality of the textual description at the website of resource owner and the resource ranking algorithm of the text-based Internet search engine. As usual, users get many irrelevant links together with good ones.

The only way to obtain relevant search results is using content-based requests. Today Recogmission LLC has launched a novel indexing and search engine Picollator®, www.picollator.com, which finds pictures and corresponding resources by image content itself. To find a resource, user submits sample image or text to Picollator.com engine. Search result depends exactly on the content itself, not necessarily on the textual description.

Instead of searching by keywords Picollator looks right inside the image and analyses detected objects. Important feature of the system is ability to find sources with pictures containing similar human faces. Recogmission search makes possible to find sites with pictures containing particular person or similar persons. Key advantage of the implemented technology is the ability to process all faces in the submitted picture and find images from the database containing those faces together or separately, as if it is a phrase.

Existing search engines match the words in text queries and tags that describe images. However, even if an image contains the appropriate keywords, it does not mean that users obtain desired search results. Therefore, it is a new approach to search results ranking. It means that user can search through large volumes of multimedia data using natural requests like any samples of visual information. The query can be described by the sentence: find web pages with pictures similar to this one.

The service is language-independent. Now it does not matter that users can not find the words to create textual search query. Picollator is able to find pictures with people wherever they are now according to the principles of the nature.

###