Search engines have emerged as one of the most useful services of the present computing age. Today’s search engines have grown into highly efficient directories and indices of information that are particularly well-suited to archiving and indexing text-based data. These services are being extended further into image and video data. Recent services like Google Goggles are extending the capabilities of search engines to visual media, but video search methods are still in their infancy and often cumbersome to use.
Current video search relies on text tags and audio track transcription methods to catalog and index objects and images within video sources. Because these methods rely on human efforts to tag all the salient features in each video, they are inherently time and resource intensive and can be prone to error and inefficiencies.
EuclidVision's feature and object modeling techniques help to identify higher-level structure in the video that can serve as the basis for more efficient and robust video search services.
Euclid Discoveries has filed a patent application for Object Archival Systems and Methods which protects the application of EuclidVision in terms of these archival and indexing operations. EuclidVision’s advanced modeling techniques and processing tools can support the automated indexing of video data, creation of object model libraries, and customization to video search.
Through use over time, these object models can grow into highly sophisticated representations of the user and visual content important to the user. The models can follow a user over networks and devices, allowing personal content to be easily accessible. Further, the models can be shared with other users, resulting in network effects which can enhance multiple video experiences for all users.
The distribution of models and model libraries can proliferate over networks and devices, being used as an advanced caching mechanism for the distribution of content. Libraries of models can be mined and models can be further analyzed and refined.