The increasing popularity of search engines such as Yahoo and Google indicates a significant demand for mechanisms of digital information retrieval. Unfortunately, current retrieval technologies are limited to textual information since a generally-recognized description of video content does not exist at this time. This technology is a system for describing video media in order to enable efficient categorization, storage, and searching functionality. The system is based on the MPEG-7 description standard developed by the International Organization for Standardization and the International Electrotechnical Commission. Each video object is defined by a number of general features which are organized into hierarchy of classes. These features can be used to build relationships between video objects and may be accessed by computer applications such as search engines, filters, and archive systems for subsequent processing.
In many settings, it can be difficult to organize, compare, or search a large collection of videos due to the varied hardware and software used for video acquisition. For example, within the same hospital, several machines from different vendors may be used to generate diagnostic videos of patients. Since the formatting and descriptions of these videos may differ, it can be difficult to manage these videos using a single database and software. This technology aims to provide a standardized description for all videos regardless of vendor, allowing for processing algorithms to operate on a homogenous set of descriptors.
A prototype of this technology has been applied to clinical echocardiography videos. Descriptors of each frame within the video were generated and used to display relevant frames to a cardiologist for observation. This technology eliminates the need for cardiologists to manually identify relevant frames and allows the cardiologist to quickly browse through multiple patients.
Patent Issued (US 8370869)
Tech Ventures Reference: IR MS99/04/08B