Picollator web search engine, the pioneering system using visual queries for image search, now equipped with a unique functionality, which is a combined cross-content search by various data types.
From its origin, the engine has developed to be language independent. The first version based on objects detection and pattern recognition technologies processed sample images and made search for similar portraits on the Web. Therefore, users were not asked for words if they had a sample photo. The new version now finds images and web resources using multi-content requests.
Picollator is going to change traditional approach to web search. It is capable to process multimedia request, including visual queries and text, to get relevant search results according to any type of the query. The system looks right inside the content and understands its essence using both pattern recognition technology and text indexing. Search results contain web pages and images with maximum relevance. It is a much more complex task – and Picollator does it.
The ranking is performed across all resources found depending on image-based ratios and text-based ratios together. Picollator is able to understand the visual meaning of some words. It does not mean that it just count the words attached to the pictures or web pages. Its engine also uses some estimation of how text can fit visual similarity of known objects.
The engine is attractive for the users having unsuccessful experience of search for the relevant information in the Web. Combination of different media in a single query helps to reduce time losses and improve search results. From a commercial point of view, the system is of interest for any company to solve the problem of digital assets and multimedia content management.
Beta version of the universal search engine Picollator is available at www.picollator.com.
