
		AT&T Laboratories Cambridge

DART Home Page

The DART project is concerned with managing collections of so-called digital assets. These assets include text and hypertext documents, images, audio and video recordings, and any other items of digital media which may be of value to an individual or organization. DART aims to provide the means to index, annotate, navigate, and retrieve from diverse collections of these assets.

In part, we intend to build upon the success of the Video Mail Retrieval project in which our lab participated, but while the VMR project concentrated exclusively on indexing video by means of speech recognition on the soundtrack, DART intends to advance those techniques and combine them with new techniques for content-based image retrieval, video parsing, user-interface design, and information retrieval in order to provide richer management of a wider range of digital assets.

Thus, underlying DART there is work in the following areas:

	speech recognition
	image analysis
	video parsing
	information retrieval

We have tied these research strands together in two primary applications: Shoebox, a tool for organizing, annotating, indexing, searching, and displaying collections of digital photographs; and AT&TV, a system which automatically and continuously records, segments, and indexes broadcast television, allowing users easily to find and watch programmes (or parts of programmes) which match their personal interests.