Company Profile, Products & Technologies, Target Market

Brief description of proposed R&D project:

High-scale multimedia information retrieval is still in its infancy; it is usually based only on metadata and cannot combine different types of features in a scalable way. The proposed project involves the development of searchable high-scale multimedia repositories, where data composed of images, video, spoken information, and metadata will be indexed into a unified multimedia index. This index can then be used to search for information based both on metadata and multi-feature similarity. We are interested in cooperating in the development of the indexing technology, evaluation techniques for the quality of results, and business scenarios in which this technology is applicable. If relevant, we are also interested in the applications of peer-2-peer technology where the searchable index is distributed across different peers, each responsible for a different type of feature and/or a particular partition of the data.

Desired Profile of R&D Partner and its role in the proposed R&D project:

We look for a partner that may have a large collection of images/video/speech and would need a solution for search in such collections. While existing solution searches only in associated metadata (e.g. by the title of a video) our solution can search by the content itself (e.g. find video clips where a given image appear). We look for partners that may help us test and validate the technology and that can help us find business cases for such technology.

Another kind of partner can be a technology company that specialize in image/video processing. While our technology can be used to index and search using low level image features we don't have the expertise to extract those features. So a partner that has such a technology may help us improve the effectiveness of our search.