Content Addressable Graphics, Image and Video Retrieval Systems

This proposal addresses development of a distributed image oriented storage and indexing architecture for building content searchable image databases (containing test, graphics, images, and digitize video). This is particularly difficult because computer vision has yet to solve the shape recognition problem which encompasses both text type and object type entities. Also, since substantial computation must be done to match image contents, substantial responsibility for the process must be delegated to the image archival engine (or in fact engines, because balancing data stored with computational elements is very important in this application as compared to more traditional text-oriented indexing and searching). In Cybernet's Phase I effort we will prototype a single node of the search engine and associated browser. Later we will scale the prototype product up for full scale Navy use.